{ "architectures": [ "DiffusionGemmaForBlockDiffusion" ], "boi_token_id": 255999, "canvas_length": 256, "dtype": "bfloat16", "eoi_token_id": 258882, "eos_token_id": [ 1, 106 ], "image_token_id": 258880, "initializer_range": 0.02, "model_type": "diffusion_gemma", "quantization_config": { "config_groups": { "group_0": { "format": "pack-quantized", "input_activations": null, "output_activations": null, "targets": [ "Linear" ], "weights": { "actorder": null, "block_structure": null, "dynamic": false, "group_size": 64, "num_bits": 4, "observer": "memoryless_minmax", "observer_kwargs": {}, "scale_dtype": null, "strategy": "group", "symmetric": true, "type": "int", "zp_dtype": null } } }, "format": "pack-quantized", "global_compression_ratio": null, "ignore": [ "model.encoder.language_model.layers.0.router", "model.encoder.language_model.layers.0.router.proj", "model.encoder.language_model.layers.0.experts.0", "model.encoder.language_model.layers.0.experts.1", "model.encoder.language_model.layers.0.experts.2", "model.encoder.language_model.layers.0.experts.3", "model.encoder.language_model.layers.0.experts.4", "model.encoder.language_model.layers.0.experts.5", "model.encoder.language_model.layers.0.experts.6", "model.encoder.language_model.layers.0.experts.7", "model.encoder.language_model.layers.0.experts.8", "model.encoder.language_model.layers.0.experts.9", "model.encoder.language_model.layers.0.experts.10", "model.encoder.language_model.layers.0.experts.11", "model.encoder.language_model.layers.0.experts.12", "model.encoder.language_model.layers.0.experts.13", "model.encoder.language_model.layers.0.experts.14", "model.encoder.language_model.layers.0.experts.15", "model.encoder.language_model.layers.0.experts.16", "model.encoder.language_model.layers.0.experts.17", "model.encoder.language_model.layers.0.experts.18", "model.encoder.language_model.layers.0.experts.19", "model.encoder.language_model.layers.0.experts.20", "model.encoder.language_model.layers.0.experts.21", "model.encoder.language_model.layers.0.experts.22", "model.encoder.language_model.layers.0.experts.23", "model.encoder.language_model.layers.0.experts.24", "model.encoder.language_model.layers.0.experts.25", "model.encoder.language_model.layers.0.experts.26", "model.encoder.language_model.layers.0.experts.27", "model.encoder.language_model.layers.0.experts.28", "model.encoder.language_model.layers.0.experts.29", "model.encoder.language_model.layers.0.experts.30", "model.encoder.language_model.layers.0.experts.31", "model.encoder.language_model.layers.0.experts.32", "model.encoder.language_model.layers.0.experts.33", "model.encoder.language_model.layers.0.experts.34", "model.encoder.language_model.layers.0.experts.35", "model.encoder.language_model.layers.0.experts.36", "model.encoder.language_model.layers.0.experts.37", "model.encoder.language_model.layers.0.experts.38", "model.encoder.language_model.layers.0.experts.39", "model.encoder.language_model.layers.0.experts.40", "model.encoder.language_model.layers.0.experts.41", "model.encoder.language_model.layers.0.experts.42", "model.encoder.language_model.layers.0.experts.43", "model.encoder.language_model.layers.0.experts.44", "model.encoder.language_model.layers.0.experts.45", "model.encoder.language_model.layers.0.experts.46", "model.encoder.language_model.layers.0.experts.47", "model.encoder.language_model.layers.0.experts.48", "model.encoder.language_model.layers.0.experts.49", "model.encoder.language_model.layers.0.experts.50", "model.encoder.language_model.layers.0.experts.51", "model.encoder.language_model.layers.0.experts.52", "model.encoder.language_model.layers.0.experts.53", "model.encoder.language_model.layers.0.experts.54", "model.encoder.language_model.layers.0.experts.55", "model.encoder.language_model.layers.0.experts.56", "model.encoder.language_model.layers.0.experts.57", "model.encoder.language_model.layers.0.experts.58", "model.encoder.language_model.layers.0.experts.59", "model.encoder.language_model.layers.0.experts.60", "model.encoder.language_model.layers.0.experts.61", "model.encoder.language_model.layers.0.experts.62", "model.encoder.language_model.layers.0.experts.63", "model.encoder.language_model.layers.0.experts.64", "model.encoder.language_model.layers.0.experts.65", "model.encoder.language_model.layers.0.experts.66", "model.encoder.language_model.layers.0.experts.67", "model.encoder.language_model.layers.0.experts.68", "model.encoder.language_model.layers.0.experts.69", "model.encoder.language_model.layers.0.experts.70", "model.encoder.language_model.layers.0.experts.71", "model.encoder.language_model.layers.0.experts.72", "model.encoder.language_model.layers.0.experts.73", "model.encoder.language_model.layers.0.experts.74", "model.encoder.language_model.layers.0.experts.75", "model.encoder.language_model.layers.0.experts.76", "model.encoder.language_model.layers.0.experts.77", "model.encoder.language_model.layers.0.experts.78", "model.encoder.language_model.layers.0.experts.79", "model.encoder.language_model.layers.0.experts.80", "model.encoder.language_model.layers.0.experts.81", "model.encoder.language_model.layers.0.experts.82", "model.encoder.language_model.layers.0.experts.83", "model.encoder.language_model.layers.0.experts.84", "model.encoder.language_model.layers.0.experts.85", "model.encoder.language_model.layers.0.experts.86", "model.encoder.language_model.layers.0.experts.87", "model.encoder.language_model.layers.0.experts.88", "model.encoder.language_model.layers.0.experts.89", "model.encoder.language_model.layers.0.experts.90", "model.encoder.language_model.layers.0.experts.91", "model.encoder.language_model.layers.0.experts.92", "model.encoder.language_model.layers.0.experts.93", "model.encoder.language_model.layers.0.experts.94", "model.encoder.language_model.layers.0.experts.95", "model.encoder.language_model.layers.0.experts.96", "model.encoder.language_model.layers.0.experts.97", "model.encoder.language_model.layers.0.experts.98", "model.encoder.language_model.layers.0.experts.99", "model.encoder.language_model.layers.0.experts.100", "model.encoder.language_model.layers.0.experts.101", "model.encoder.language_model.layers.0.experts.102", "model.encoder.language_model.layers.0.experts.103", "model.encoder.language_model.layers.0.experts.104", "model.encoder.language_model.layers.0.experts.105", "model.encoder.language_model.layers.0.experts.106", "model.encoder.language_model.layers.0.experts.107", "model.encoder.language_model.layers.0.experts.108", "model.encoder.language_model.layers.0.experts.109", "model.encoder.language_model.layers.0.experts.110", "model.encoder.language_model.layers.0.experts.111", "model.encoder.language_model.layers.0.experts.112", "model.encoder.language_model.layers.0.experts.113", "model.encoder.language_model.layers.0.experts.114", "model.encoder.language_model.layers.0.experts.115", "model.encoder.language_model.layers.0.experts.116", "model.encoder.language_model.layers.0.experts.117", "model.encoder.language_model.layers.0.experts.118", "model.encoder.language_model.layers.0.experts.119", "model.encoder.language_model.layers.0.experts.120", "model.encoder.language_model.layers.0.experts.121", "model.encoder.language_model.layers.0.experts.122", "model.encoder.language_model.layers.0.experts.123", "model.encoder.language_model.layers.0.experts.124", "model.encoder.language_model.layers.0.experts.125", "model.encoder.language_model.layers.0.experts.126", "model.encoder.language_model.layers.0.experts.127", "model.encoder.language_model.layers.1.router", "model.encoder.language_model.layers.1.router.proj", "model.encoder.language_model.layers.1.experts.0", "model.encoder.language_model.layers.1.experts.1", "model.encoder.language_model.layers.1.experts.2", "model.encoder.language_model.layers.1.experts.3", "model.encoder.language_model.layers.1.experts.4", "model.encoder.language_model.layers.1.experts.5", "model.encoder.language_model.layers.1.experts.6", "model.encoder.language_model.layers.1.experts.7", "model.encoder.language_model.layers.1.experts.8", "model.encoder.language_model.layers.1.experts.9", "model.encoder.language_model.layers.1.experts.10", "model.encoder.language_model.layers.1.experts.11", "model.encoder.language_model.layers.1.experts.12", "model.encoder.language_model.layers.1.experts.13", "model.encoder.language_model.layers.1.experts.14", "model.encoder.language_model.layers.1.experts.15", "model.encoder.language_model.layers.1.experts.16", "model.encoder.language_model.layers.1.experts.17", "model.encoder.language_model.layers.1.experts.18", "model.encoder.language_model.layers.1.experts.19", "model.encoder.language_model.layers.1.experts.20", "model.encoder.language_model.layers.1.experts.21", "model.encoder.language_model.layers.1.experts.22", "model.encoder.language_model.layers.1.experts.23", "model.encoder.language_model.layers.1.experts.24", "model.encoder.language_model.layers.1.experts.25", "model.encoder.language_model.layers.1.experts.26", "model.encoder.language_model.layers.1.experts.27", "model.encoder.language_model.layers.1.experts.28", "model.encoder.language_model.layers.1.experts.29", "model.encoder.language_model.layers.1.experts.30", "model.encoder.language_model.layers.1.experts.31", "model.encoder.language_model.layers.1.experts.32", "model.encoder.language_model.layers.1.experts.33", "model.encoder.language_model.layers.1.experts.34", "model.encoder.language_model.layers.1.experts.35", "model.encoder.language_model.layers.1.experts.36", "model.encoder.language_model.layers.1.experts.37", "model.encoder.language_model.layers.1.experts.38", "model.encoder.language_model.layers.1.experts.39", "model.encoder.language_model.layers.1.experts.40", "model.encoder.language_model.layers.1.experts.41", "model.encoder.language_model.layers.1.experts.42", "model.encoder.language_model.layers.1.experts.43", "model.encoder.language_model.layers.1.experts.44", "model.encoder.language_model.layers.1.experts.45", "model.encoder.language_model.layers.1.experts.46", "model.encoder.language_model.layers.1.experts.47", "model.encoder.language_model.layers.1.experts.48", "model.encoder.language_model.layers.1.experts.49", "model.encoder.language_model.layers.1.experts.50", "model.encoder.language_model.layers.1.experts.51", "model.encoder.language_model.layers.1.experts.52", "model.encoder.language_model.layers.1.experts.53", "model.encoder.language_model.layers.1.experts.54", "model.encoder.language_model.layers.1.experts.55", "model.encoder.language_model.layers.1.experts.56", "model.encoder.language_model.layers.1.experts.57", "model.encoder.language_model.layers.1.experts.58", "model.encoder.language_model.layers.1.experts.59", "model.encoder.language_model.layers.1.experts.60", "model.encoder.language_model.layers.1.experts.61", "model.encoder.language_model.layers.1.experts.62", "model.encoder.language_model.layers.1.experts.63", "model.encoder.language_model.layers.1.experts.64", "model.encoder.language_model.layers.1.experts.65", "model.encoder.language_model.layers.1.experts.66", "model.encoder.language_model.layers.1.experts.67", "model.encoder.language_model.layers.1.experts.68", "model.encoder.language_model.layers.1.experts.69", "model.encoder.language_model.layers.1.experts.70", "model.encoder.language_model.layers.1.experts.71", "model.encoder.language_model.layers.1.experts.72", "model.encoder.language_model.layers.1.experts.73", "model.encoder.language_model.layers.1.experts.74", "model.encoder.language_model.layers.1.experts.75", "model.encoder.language_model.layers.1.experts.76", "model.encoder.language_model.layers.1.experts.77", "model.encoder.language_model.layers.1.experts.78", "model.encoder.language_model.layers.1.experts.79", "model.encoder.language_model.layers.1.experts.80", "model.encoder.language_model.layers.1.experts.81", "model.encoder.language_model.layers.1.experts.82", "model.encoder.language_model.layers.1.experts.83", "model.encoder.language_model.layers.1.experts.84", "model.encoder.language_model.layers.1.experts.85", "model.encoder.language_model.layers.1.experts.86", "model.encoder.language_model.layers.1.experts.87", "model.encoder.language_model.layers.1.experts.88", "model.encoder.language_model.layers.1.experts.89", "model.encoder.language_model.layers.1.experts.90", "model.encoder.language_model.layers.1.experts.91", "model.encoder.language_model.layers.1.experts.92", "model.encoder.language_model.layers.1.experts.93", "model.encoder.language_model.layers.1.experts.94", "model.encoder.language_model.layers.1.experts.95", "model.encoder.language_model.layers.1.experts.96", "model.encoder.language_model.layers.1.experts.97", "model.encoder.language_model.layers.1.experts.98", "model.encoder.language_model.layers.1.experts.99", "model.encoder.language_model.layers.1.experts.100", "model.encoder.language_model.layers.1.experts.101", "model.encoder.language_model.layers.1.experts.102", "model.encoder.language_model.layers.1.experts.103", "model.encoder.language_model.layers.1.experts.104", "model.encoder.language_model.layers.1.experts.105", "model.encoder.language_model.layers.1.experts.106", "model.encoder.language_model.layers.1.experts.107", "model.encoder.language_model.layers.1.experts.108", "model.encoder.language_model.layers.1.experts.109", "model.encoder.language_model.layers.1.experts.110", "model.encoder.language_model.layers.1.experts.111", "model.encoder.language_model.layers.1.experts.112", "model.encoder.language_model.layers.1.experts.113", "model.encoder.language_model.layers.1.experts.114", "model.encoder.language_model.layers.1.experts.115", "model.encoder.language_model.layers.1.experts.116", "model.encoder.language_model.layers.1.experts.117", "model.encoder.language_model.layers.1.experts.118", "model.encoder.language_model.layers.1.experts.119", "model.encoder.language_model.layers.1.experts.120", "model.encoder.language_model.layers.1.experts.121", "model.encoder.language_model.layers.1.experts.122", "model.encoder.language_model.layers.1.experts.123", "model.encoder.language_model.layers.1.experts.124", "model.encoder.language_model.layers.1.experts.125", "model.encoder.language_model.layers.1.experts.126", "model.encoder.language_model.layers.1.experts.127", "model.encoder.language_model.layers.2.router", "model.encoder.language_model.layers.2.router.proj", "model.encoder.language_model.layers.2.experts.0", "model.encoder.language_model.layers.2.experts.1", "model.encoder.language_model.layers.2.experts.2", "model.encoder.language_model.layers.2.experts.3", "model.encoder.language_model.layers.2.experts.4", "model.encoder.language_model.layers.2.experts.5", "model.encoder.language_model.layers.2.experts.6", "model.encoder.language_model.layers.2.experts.7", "model.encoder.language_model.layers.2.experts.8", "model.encoder.language_model.layers.2.experts.9", "model.encoder.language_model.layers.2.experts.10", "model.encoder.language_model.layers.2.experts.11", "model.encoder.language_model.layers.2.experts.12", "model.encoder.language_model.layers.2.experts.13", "model.encoder.language_model.layers.2.experts.14", "model.encoder.language_model.layers.2.experts.15", "model.encoder.language_model.layers.2.experts.16", "model.encoder.language_model.layers.2.experts.17", "model.encoder.language_model.layers.2.experts.18", "model.encoder.language_model.layers.2.experts.19", "model.encoder.language_model.layers.2.experts.20", "model.encoder.language_model.layers.2.experts.21", "model.encoder.language_model.layers.2.experts.22", "model.encoder.language_model.layers.2.experts.23", "model.encoder.language_model.layers.2.experts.24", "model.encoder.language_model.layers.2.experts.25", "model.encoder.language_model.layers.2.experts.26", "model.encoder.language_model.layers.2.experts.27", "model.encoder.language_model.layers.2.experts.28", "model.encoder.language_model.layers.2.experts.29", "model.encoder.language_model.layers.2.experts.30", "model.encoder.language_model.layers.2.experts.31", "model.encoder.language_model.layers.2.experts.32", "model.encoder.language_model.layers.2.experts.33", "model.encoder.language_model.layers.2.experts.34", "model.encoder.language_model.layers.2.experts.35", "model.encoder.language_model.layers.2.experts.36", "model.encoder.language_model.layers.2.experts.37", "model.encoder.language_model.layers.2.experts.38", "model.encoder.language_model.layers.2.experts.39", "model.encoder.language_model.layers.2.experts.40", "model.encoder.language_model.layers.2.experts.41", "model.encoder.language_model.layers.2.experts.42", "model.encoder.language_model.layers.2.experts.43", "model.encoder.language_model.layers.2.experts.44", "model.encoder.language_model.layers.2.experts.45", "model.encoder.language_model.layers.2.experts.46", "model.encoder.language_model.layers.2.experts.47", "model.encoder.language_model.layers.2.experts.48", "model.encoder.language_model.layers.2.experts.49", "model.encoder.language_model.layers.2.experts.50", "model.encoder.language_model.layers.2.experts.51", "model.encoder.language_model.layers.2.experts.52", "model.encoder.language_model.layers.2.experts.53", "model.encoder.language_model.layers.2.experts.54", "model.encoder.language_model.layers.2.experts.55", "model.encoder.language_model.layers.2.experts.56", "model.encoder.language_model.layers.2.experts.57", "model.encoder.language_model.layers.2.experts.58", "model.encoder.language_model.layers.2.experts.59", "model.encoder.language_model.layers.2.experts.60", "model.encoder.language_model.layers.2.experts.61", "model.encoder.language_model.layers.2.experts.62", "model.encoder.language_model.layers.2.experts.63", "model.encoder.language_model.layers.2.experts.64", "model.encoder.language_model.layers.2.experts.65", "model.encoder.language_model.layers.2.experts.66", "model.encoder.language_model.layers.2.experts.67", "model.encoder.language_model.layers.2.experts.68", "model.encoder.language_model.layers.2.experts.69", "model.encoder.language_model.layers.2.experts.70", "model.encoder.language_model.layers.2.experts.71", "model.encoder.language_model.layers.2.experts.72", "model.encoder.language_model.layers.2.experts.73", "model.encoder.language_model.layers.2.experts.74", "model.encoder.language_model.layers.2.experts.75", "model.encoder.language_model.layers.2.experts.76", "model.encoder.language_model.layers.2.experts.77", "model.encoder.language_model.layers.2.experts.78", "model.encoder.language_model.layers.2.experts.79", "model.encoder.language_model.layers.2.experts.80", "model.encoder.language_model.layers.2.experts.81", "model.encoder.language_model.layers.2.experts.82", "model.encoder.language_model.layers.2.experts.83", "model.encoder.language_model.layers.2.experts.84", "model.encoder.language_model.layers.2.experts.85", "model.encoder.language_model.layers.2.experts.86", "model.encoder.language_model.layers.2.experts.87", "model.encoder.language_model.layers.2.experts.88", "model.encoder.language_model.layers.2.experts.89", "model.encoder.language_model.layers.2.experts.90", "model.encoder.language_model.layers.2.experts.91", "model.encoder.language_model.layers.2.experts.92", "model.encoder.language_model.layers.2.experts.93", "model.encoder.language_model.layers.2.experts.94", "model.encoder.language_model.layers.2.experts.95", "model.encoder.language_model.layers.2.experts.96", "model.encoder.language_model.layers.2.experts.97", "model.encoder.language_model.layers.2.experts.98", "model.encoder.language_model.layers.2.experts.99", "model.encoder.language_model.layers.2.experts.100", "model.encoder.language_model.layers.2.experts.101", "model.encoder.language_model.layers.2.experts.102", "model.encoder.language_model.layers.2.experts.103", "model.encoder.language_model.layers.2.experts.104", "model.encoder.language_model.layers.2.experts.105", "model.encoder.language_model.layers.2.experts.106", "model.encoder.language_model.layers.2.experts.107", "model.encoder.language_model.layers.2.experts.108", "model.encoder.language_model.layers.2.experts.109", "model.encoder.language_model.layers.2.experts.110", "model.encoder.language_model.layers.2.experts.111", "model.encoder.language_model.layers.2.experts.112", "model.encoder.language_model.layers.2.experts.113", "model.encoder.language_model.layers.2.experts.114", "model.encoder.language_model.layers.2.experts.115", "model.encoder.language_model.layers.2.experts.116", "model.encoder.language_model.layers.2.experts.117", "model.encoder.language_model.layers.2.experts.118", "model.encoder.language_model.layers.2.experts.119", "model.encoder.language_model.layers.2.experts.120", "model.encoder.language_model.layers.2.experts.121", "model.encoder.language_model.layers.2.experts.122", "model.encoder.language_model.layers.2.experts.123", "model.encoder.language_model.layers.2.experts.124", "model.encoder.language_model.layers.2.experts.125", "model.encoder.language_model.layers.2.experts.126", "model.encoder.language_model.layers.2.experts.127", "model.encoder.language_model.layers.3.router", "model.encoder.language_model.layers.3.router.proj", "model.encoder.language_model.layers.3.experts.0", "model.encoder.language_model.layers.3.experts.1", "model.encoder.language_model.layers.3.experts.2", "model.encoder.language_model.layers.3.experts.3", "model.encoder.language_model.layers.3.experts.4", "model.encoder.language_model.layers.3.experts.5", "model.encoder.language_model.layers.3.experts.6", "model.encoder.language_model.layers.3.experts.7", "model.encoder.language_model.layers.3.experts.8", "model.encoder.language_model.layers.3.experts.9", "model.encoder.language_model.layers.3.experts.10", "model.encoder.language_model.layers.3.experts.11", "model.encoder.language_model.layers.3.experts.12", "model.encoder.language_model.layers.3.experts.13", "model.encoder.language_model.layers.3.experts.14", "model.encoder.language_model.layers.3.experts.15", "model.encoder.language_model.layers.3.experts.16", "model.encoder.language_model.layers.3.experts.17", "model.encoder.language_model.layers.3.experts.18", "model.encoder.language_model.layers.3.experts.19", "model.encoder.language_model.layers.3.experts.20", "model.encoder.language_model.layers.3.experts.21", "model.encoder.language_model.layers.3.experts.22", "model.encoder.language_model.layers.3.experts.23", "model.encoder.language_model.layers.3.experts.24", "model.encoder.language_model.layers.3.experts.25", "model.encoder.language_model.layers.3.experts.26", "model.encoder.language_model.layers.3.experts.27", "model.encoder.language_model.layers.3.experts.28", "model.encoder.language_model.layers.3.experts.29", "model.encoder.language_model.layers.3.experts.30", "model.encoder.language_model.layers.3.experts.31", "model.encoder.language_model.layers.3.experts.32", "model.encoder.language_model.layers.3.experts.33", "model.encoder.language_model.layers.3.experts.34", "model.encoder.language_model.layers.3.experts.35", "model.encoder.language_model.layers.3.experts.36", "model.encoder.language_model.layers.3.experts.37", "model.encoder.language_model.layers.3.experts.38", "model.encoder.language_model.layers.3.experts.39", "model.encoder.language_model.layers.3.experts.40", "model.encoder.language_model.layers.3.experts.41", "model.encoder.language_model.layers.3.experts.42", "model.encoder.language_model.layers.3.experts.43", "model.encoder.language_model.layers.3.experts.44", "model.encoder.language_model.layers.3.experts.45", "model.encoder.language_model.layers.3.experts.46", "model.encoder.language_model.layers.3.experts.47", "model.encoder.language_model.layers.3.experts.48", "model.encoder.language_model.layers.3.experts.49", "model.encoder.language_model.layers.3.experts.50", "model.encoder.language_model.layers.3.experts.51", "model.encoder.language_model.layers.3.experts.52", "model.encoder.language_model.layers.3.experts.53", "model.encoder.language_model.layers.3.experts.54", "model.encoder.language_model.layers.3.experts.55", "model.encoder.language_model.layers.3.experts.56", "model.encoder.language_model.layers.3.experts.57", "model.encoder.language_model.layers.3.experts.58", "model.encoder.language_model.layers.3.experts.59", "model.encoder.language_model.layers.3.experts.60", "model.encoder.language_model.layers.3.experts.61", "model.encoder.language_model.layers.3.experts.62", "model.encoder.language_model.layers.3.experts.63", "model.encoder.language_model.layers.3.experts.64", "model.encoder.language_model.layers.3.experts.65", "model.encoder.language_model.layers.3.experts.66", "model.encoder.language_model.layers.3.experts.67", "model.encoder.language_model.layers.3.experts.68", "model.encoder.language_model.layers.3.experts.69", "model.encoder.language_model.layers.3.experts.70", "model.encoder.language_model.layers.3.experts.71", "model.encoder.language_model.layers.3.experts.72", "model.encoder.language_model.layers.3.experts.73", "model.encoder.language_model.layers.3.experts.74", "model.encoder.language_model.layers.3.experts.75", "model.encoder.language_model.layers.3.experts.76", "model.encoder.language_model.layers.3.experts.77", "model.encoder.language_model.layers.3.experts.78", "model.encoder.language_model.layers.3.experts.79", "model.encoder.language_model.layers.3.experts.80", "model.encoder.language_model.layers.3.experts.81", "model.encoder.language_model.layers.3.experts.82", "model.encoder.language_model.layers.3.experts.83", "model.encoder.language_model.layers.3.experts.84", "model.encoder.language_model.layers.3.experts.85", "model.encoder.language_model.layers.3.experts.86", "model.encoder.language_model.layers.3.experts.87", "model.encoder.language_model.layers.3.experts.88", "model.encoder.language_model.layers.3.experts.89", "model.encoder.language_model.layers.3.experts.90", "model.encoder.language_model.layers.3.experts.91", "model.encoder.language_model.layers.3.experts.92", "model.encoder.language_model.layers.3.experts.93", "model.encoder.language_model.layers.3.experts.94", "model.encoder.language_model.layers.3.experts.95", "model.encoder.language_model.layers.3.experts.96", "model.encoder.language_model.layers.3.experts.97", "model.encoder.language_model.layers.3.experts.98", "model.encoder.language_model.layers.3.experts.99", "model.encoder.language_model.layers.3.experts.100", "model.encoder.language_model.layers.3.experts.101", "model.encoder.language_model.layers.3.experts.102", "model.encoder.language_model.layers.3.experts.103", "model.encoder.language_model.layers.3.experts.104", "model.encoder.language_model.layers.3.experts.105", "model.encoder.language_model.layers.3.experts.106", "model.encoder.language_model.layers.3.experts.107", "model.encoder.language_model.layers.3.experts.108", "model.encoder.language_model.layers.3.experts.109", "model.encoder.language_model.layers.3.experts.110", "model.encoder.language_model.layers.3.experts.111", "model.encoder.language_model.layers.3.experts.112", "model.encoder.language_model.layers.3.experts.113", "model.encoder.language_model.layers.3.experts.114", "model.encoder.language_model.layers.3.experts.115", "model.encoder.language_model.layers.3.experts.116", "model.encoder.language_model.layers.3.experts.117", "model.encoder.language_model.layers.3.experts.118", "model.encoder.language_model.layers.3.experts.119", "model.encoder.language_model.layers.3.experts.120", "model.encoder.language_model.layers.3.experts.121", "model.encoder.language_model.layers.3.experts.122", "model.encoder.language_model.layers.3.experts.123", "model.encoder.language_model.layers.3.experts.124", "model.encoder.language_model.layers.3.experts.125", "model.encoder.language_model.layers.3.experts.126", "model.encoder.language_model.layers.3.experts.127", "model.encoder.language_model.layers.4.router", "model.encoder.language_model.layers.4.router.proj", "model.encoder.language_model.layers.4.experts.0", "model.encoder.language_model.layers.4.experts.1", "model.encoder.language_model.layers.4.experts.2", "model.encoder.language_model.layers.4.experts.3", "model.encoder.language_model.layers.4.experts.4", "model.encoder.language_model.layers.4.experts.5", "model.encoder.language_model.layers.4.experts.6", "model.encoder.language_model.layers.4.experts.7", "model.encoder.language_model.layers.4.experts.8", "model.encoder.language_model.layers.4.experts.9", "model.encoder.language_model.layers.4.experts.10", "model.encoder.language_model.layers.4.experts.11", "model.encoder.language_model.layers.4.experts.12", "model.encoder.language_model.layers.4.experts.13", "model.encoder.language_model.layers.4.experts.14", "model.encoder.language_model.layers.4.experts.15", "model.encoder.language_model.layers.4.experts.16", "model.encoder.language_model.layers.4.experts.17", "model.encoder.language_model.layers.4.experts.18", "model.encoder.language_model.layers.4.experts.19", "model.encoder.language_model.layers.4.experts.20", "model.encoder.language_model.layers.4.experts.21", "model.encoder.language_model.layers.4.experts.22", "model.encoder.language_model.layers.4.experts.23", "model.encoder.language_model.layers.4.experts.24", "model.encoder.language_model.layers.4.experts.25", "model.encoder.language_model.layers.4.experts.26", "model.encoder.language_model.layers.4.experts.27", "model.encoder.language_model.layers.4.experts.28", "model.encoder.language_model.layers.4.experts.29", "model.encoder.language_model.layers.4.experts.30", "model.encoder.language_model.layers.4.experts.31", "model.encoder.language_model.layers.4.experts.32", "model.encoder.language_model.layers.4.experts.33", "model.encoder.language_model.layers.4.experts.34", "model.encoder.language_model.layers.4.experts.35", "model.encoder.language_model.layers.4.experts.36", "model.encoder.language_model.layers.4.experts.37", "model.encoder.language_model.layers.4.experts.38", "model.encoder.language_model.layers.4.experts.39", "model.encoder.language_model.layers.4.experts.40", "model.encoder.language_model.layers.4.experts.41", "model.encoder.language_model.layers.4.experts.42", "model.encoder.language_model.layers.4.experts.43", "model.encoder.language_model.layers.4.experts.44", "model.encoder.language_model.layers.4.experts.45", "model.encoder.language_model.layers.4.experts.46", "model.encoder.language_model.layers.4.experts.47", "model.encoder.language_model.layers.4.experts.48", "model.encoder.language_model.layers.4.experts.49", "model.encoder.language_model.layers.4.experts.50", "model.encoder.language_model.layers.4.experts.51", "model.encoder.language_model.layers.4.experts.52", "model.encoder.language_model.layers.4.experts.53", "model.encoder.language_model.layers.4.experts.54", "model.encoder.language_model.layers.4.experts.55", "model.encoder.language_model.layers.4.experts.56", "model.encoder.language_model.layers.4.experts.57", "model.encoder.language_model.layers.4.experts.58", "model.encoder.language_model.layers.4.experts.59", "model.encoder.language_model.layers.4.experts.60", "model.encoder.language_model.layers.4.experts.61", "model.encoder.language_model.layers.4.experts.62", "model.encoder.language_model.layers.4.experts.63", "model.encoder.language_model.layers.4.experts.64", "model.encoder.language_model.layers.4.experts.65", "model.encoder.language_model.layers.4.experts.66", "model.encoder.language_model.layers.4.experts.67", "model.encoder.language_model.layers.4.experts.68", "model.encoder.language_model.layers.4.experts.69", "model.encoder.language_model.layers.4.experts.70", "model.encoder.language_model.layers.4.experts.71", "model.encoder.language_model.layers.4.experts.72", "model.encoder.language_model.layers.4.experts.73", "model.encoder.language_model.layers.4.experts.74", "model.encoder.language_model.layers.4.experts.75", "model.encoder.language_model.layers.4.experts.76", "model.encoder.language_model.layers.4.experts.77", "model.encoder.language_model.layers.4.experts.78", "model.encoder.language_model.layers.4.experts.79", "model.encoder.language_model.layers.4.experts.80", "model.encoder.language_model.layers.4.experts.81", "model.encoder.language_model.layers.4.experts.82", "model.encoder.language_model.layers.4.experts.83", "model.encoder.language_model.layers.4.experts.84", "model.encoder.language_model.layers.4.experts.85", "model.encoder.language_model.layers.4.experts.86", "model.encoder.language_model.layers.4.experts.87", "model.encoder.language_model.layers.4.experts.88", "model.encoder.language_model.layers.4.experts.89", "model.encoder.language_model.layers.4.experts.90", "model.encoder.language_model.layers.4.experts.91", "model.encoder.language_model.layers.4.experts.92", "model.encoder.language_model.layers.4.experts.93", "model.encoder.language_model.layers.4.experts.94", "model.encoder.language_model.layers.4.experts.95", "model.encoder.language_model.layers.4.experts.96", "model.encoder.language_model.layers.4.experts.97", "model.encoder.language_model.layers.4.experts.98", "model.encoder.language_model.layers.4.experts.99", "model.encoder.language_model.layers.4.experts.100", "model.encoder.language_model.layers.4.experts.101", "model.encoder.language_model.layers.4.experts.102", "model.encoder.language_model.layers.4.experts.103", "model.encoder.language_model.layers.4.experts.104", "model.encoder.language_model.layers.4.experts.105", "model.encoder.language_model.layers.4.experts.106", "model.encoder.language_model.layers.4.experts.107", "model.encoder.language_model.layers.4.experts.108", "model.encoder.language_model.layers.4.experts.109", "model.encoder.language_model.layers.4.experts.110", "model.encoder.language_model.layers.4.experts.111", "model.encoder.language_model.layers.4.experts.112", "model.encoder.language_model.layers.4.experts.113", "model.encoder.language_model.layers.4.experts.114", "model.encoder.language_model.layers.4.experts.115", "model.encoder.language_model.layers.4.experts.116", "model.encoder.language_model.layers.4.experts.117", "model.encoder.language_model.layers.4.experts.118", "model.encoder.language_model.layers.4.experts.119", "model.encoder.language_model.layers.4.experts.120", "model.encoder.language_model.layers.4.experts.121", "model.encoder.language_model.layers.4.experts.122", "model.encoder.language_model.layers.4.experts.123", "model.encoder.language_model.layers.4.experts.124", "model.encoder.language_model.layers.4.experts.125", "model.encoder.language_model.layers.4.experts.126", "model.encoder.language_model.layers.4.experts.127", "model.encoder.language_model.layers.5.router", "model.encoder.language_model.layers.5.router.proj", "model.encoder.language_model.layers.5.experts.0", "model.encoder.language_model.layers.5.experts.1", "model.encoder.language_model.layers.5.experts.2", "model.encoder.language_model.layers.5.experts.3", "model.encoder.language_model.layers.5.experts.4", "model.encoder.language_model.layers.5.experts.5", "model.encoder.language_model.layers.5.experts.6", "model.encoder.language_model.layers.5.experts.7", "model.encoder.language_model.layers.5.experts.8", "model.encoder.language_model.layers.5.experts.9", "model.encoder.language_model.layers.5.experts.10", "model.encoder.language_model.layers.5.experts.11", "model.encoder.language_model.layers.5.experts.12", "model.encoder.language_model.layers.5.experts.13", "model.encoder.language_model.layers.5.experts.14", "model.encoder.language_model.layers.5.experts.15", "model.encoder.language_model.layers.5.experts.16", "model.encoder.language_model.layers.5.experts.17", "model.encoder.language_model.layers.5.experts.18", "model.encoder.language_model.layers.5.experts.19", "model.encoder.language_model.layers.5.experts.20", "model.encoder.language_model.layers.5.experts.21", "model.encoder.language_model.layers.5.experts.22", "model.encoder.language_model.layers.5.experts.23", "model.encoder.language_model.layers.5.experts.24", "model.encoder.language_model.layers.5.experts.25", "model.encoder.language_model.layers.5.experts.26", "model.encoder.language_model.layers.5.experts.27", "model.encoder.language_model.layers.5.experts.28", "model.encoder.language_model.layers.5.experts.29", "model.encoder.language_model.layers.5.experts.30", "model.encoder.language_model.layers.5.experts.31", "model.encoder.language_model.layers.5.experts.32", "model.encoder.language_model.layers.5.experts.33", "model.encoder.language_model.layers.5.experts.34", "model.encoder.language_model.layers.5.experts.35", "model.encoder.language_model.layers.5.experts.36", "model.encoder.language_model.layers.5.experts.37", "model.encoder.language_model.layers.5.experts.38", "model.encoder.language_model.layers.5.experts.39", "model.encoder.language_model.layers.5.experts.40", "model.encoder.language_model.layers.5.experts.41", "model.encoder.language_model.layers.5.experts.42", "model.encoder.language_model.layers.5.experts.43", "model.encoder.language_model.layers.5.experts.44", "model.encoder.language_model.layers.5.experts.45", "model.encoder.language_model.layers.5.experts.46", "model.encoder.language_model.layers.5.experts.47", "model.encoder.language_model.layers.5.experts.48", "model.encoder.language_model.layers.5.experts.49", "model.encoder.language_model.layers.5.experts.50", "model.encoder.language_model.layers.5.experts.51", "model.encoder.language_model.layers.5.experts.52", "model.encoder.language_model.layers.5.experts.53", "model.encoder.language_model.layers.5.experts.54", "model.encoder.language_model.layers.5.experts.55", "model.encoder.language_model.layers.5.experts.56", "model.encoder.language_model.layers.5.experts.57", "model.encoder.language_model.layers.5.experts.58", "model.encoder.language_model.layers.5.experts.59", "model.encoder.language_model.layers.5.experts.60", "model.encoder.language_model.layers.5.experts.61", "model.encoder.language_model.layers.5.experts.62", "model.encoder.language_model.layers.5.experts.63", "model.encoder.language_model.layers.5.experts.64", "model.encoder.language_model.layers.5.experts.65", "model.encoder.language_model.layers.5.experts.66", "model.encoder.language_model.layers.5.experts.67", "model.encoder.language_model.layers.5.experts.68", "model.encoder.language_model.layers.5.experts.69", "model.encoder.language_model.layers.5.experts.70", "model.encoder.language_model.layers.5.experts.71", "model.encoder.language_model.layers.5.experts.72", "model.encoder.language_model.layers.5.experts.73", "model.encoder.language_model.layers.5.experts.74", "model.encoder.language_model.layers.5.experts.75", "model.encoder.language_model.layers.5.experts.76", "model.encoder.language_model.layers.5.experts.77", "model.encoder.language_model.layers.5.experts.78", "model.encoder.language_model.layers.5.experts.79", "model.encoder.language_model.layers.5.experts.80", "model.encoder.language_model.layers.5.experts.81", "model.encoder.language_model.layers.5.experts.82", "model.encoder.language_model.layers.5.experts.83", "model.encoder.language_model.layers.5.experts.84", "model.encoder.language_model.layers.5.experts.85", "model.encoder.language_model.layers.5.experts.86", "model.encoder.language_model.layers.5.experts.87", "model.encoder.language_model.layers.5.experts.88", "model.encoder.language_model.layers.5.experts.89", "model.encoder.language_model.layers.5.experts.90", "model.encoder.language_model.layers.5.experts.91", "model.encoder.language_model.layers.5.experts.92", "model.encoder.language_model.layers.5.experts.93", "model.encoder.language_model.layers.5.experts.94", "model.encoder.language_model.layers.5.experts.95", "model.encoder.language_model.layers.5.experts.96", "model.encoder.language_model.layers.5.experts.97", "model.encoder.language_model.layers.5.experts.98", "model.encoder.language_model.layers.5.experts.99", "model.encoder.language_model.layers.5.experts.100", "model.encoder.language_model.layers.5.experts.101", "model.encoder.language_model.layers.5.experts.102", "model.encoder.language_model.layers.5.experts.103", "model.encoder.language_model.layers.5.experts.104", "model.encoder.language_model.layers.5.experts.105", "model.encoder.language_model.layers.5.experts.106", "model.encoder.language_model.layers.5.experts.107", "model.encoder.language_model.layers.5.experts.108", "model.encoder.language_model.layers.5.experts.109", "model.encoder.language_model.layers.5.experts.110", "model.encoder.language_model.layers.5.experts.111", "model.encoder.language_model.layers.5.experts.112", "model.encoder.language_model.layers.5.experts.113", "model.encoder.language_model.layers.5.experts.114", "model.encoder.language_model.layers.5.experts.115", "model.encoder.language_model.layers.5.experts.116", "model.encoder.language_model.layers.5.experts.117", "model.encoder.language_model.layers.5.experts.118", "model.encoder.language_model.layers.5.experts.119", "model.encoder.language_model.layers.5.experts.120", "model.encoder.language_model.layers.5.experts.121", "model.encoder.language_model.layers.5.experts.122", "model.encoder.language_model.layers.5.experts.123", "model.encoder.language_model.layers.5.experts.124", "model.encoder.language_model.layers.5.experts.125", "model.encoder.language_model.layers.5.experts.126", "model.encoder.language_model.layers.5.experts.127", "model.encoder.language_model.layers.6.router", "model.encoder.language_model.layers.6.router.proj", "model.encoder.language_model.layers.6.experts.0", "model.encoder.language_model.layers.6.experts.1", "model.encoder.language_model.layers.6.experts.2", "model.encoder.language_model.layers.6.experts.3", "model.encoder.language_model.layers.6.experts.4", "model.encoder.language_model.layers.6.experts.5", "model.encoder.language_model.layers.6.experts.6", "model.encoder.language_model.layers.6.experts.7", "model.encoder.language_model.layers.6.experts.8", "model.encoder.language_model.layers.6.experts.9", "model.encoder.language_model.layers.6.experts.10", "model.encoder.language_model.layers.6.experts.11", "model.encoder.language_model.layers.6.experts.12", "model.encoder.language_model.layers.6.experts.13", "model.encoder.language_model.layers.6.experts.14", "model.encoder.language_model.layers.6.experts.15", "model.encoder.language_model.layers.6.experts.16", "model.encoder.language_model.layers.6.experts.17", "model.encoder.language_model.layers.6.experts.18", "model.encoder.language_model.layers.6.experts.19", "model.encoder.language_model.layers.6.experts.20", "model.encoder.language_model.layers.6.experts.21", "model.encoder.language_model.layers.6.experts.22", "model.encoder.language_model.layers.6.experts.23", "model.encoder.language_model.layers.6.experts.24", "model.encoder.language_model.layers.6.experts.25", "model.encoder.language_model.layers.6.experts.26", "model.encoder.language_model.layers.6.experts.27", "model.encoder.language_model.layers.6.experts.28", "model.encoder.language_model.layers.6.experts.29", "model.encoder.language_model.layers.6.experts.30", "model.encoder.language_model.layers.6.experts.31", "model.encoder.language_model.layers.6.experts.32", "model.encoder.language_model.layers.6.experts.33", "model.encoder.language_model.layers.6.experts.34", "model.encoder.language_model.layers.6.experts.35", "model.encoder.language_model.layers.6.experts.36", "model.encoder.language_model.layers.6.experts.37", "model.encoder.language_model.layers.6.experts.38", "model.encoder.language_model.layers.6.experts.39", "model.encoder.language_model.layers.6.experts.40", "model.encoder.language_model.layers.6.experts.41", "model.encoder.language_model.layers.6.experts.42", "model.encoder.language_model.layers.6.experts.43", "model.encoder.language_model.layers.6.experts.44", "model.encoder.language_model.layers.6.experts.45", "model.encoder.language_model.layers.6.experts.46", "model.encoder.language_model.layers.6.experts.47", "model.encoder.language_model.layers.6.experts.48", "model.encoder.language_model.layers.6.experts.49", "model.encoder.language_model.layers.6.experts.50", "model.encoder.language_model.layers.6.experts.51", "model.encoder.language_model.layers.6.experts.52", "model.encoder.language_model.layers.6.experts.53", "model.encoder.language_model.layers.6.experts.54", "model.encoder.language_model.layers.6.experts.55", "model.encoder.language_model.layers.6.experts.56", "model.encoder.language_model.layers.6.experts.57", "model.encoder.language_model.layers.6.experts.58", "model.encoder.language_model.layers.6.experts.59", "model.encoder.language_model.layers.6.experts.60", "model.encoder.language_model.layers.6.experts.61", "model.encoder.language_model.layers.6.experts.62", "model.encoder.language_model.layers.6.experts.63", "model.encoder.language_model.layers.6.experts.64", "model.encoder.language_model.layers.6.experts.65", "model.encoder.language_model.layers.6.experts.66", "model.encoder.language_model.layers.6.experts.67", "model.encoder.language_model.layers.6.experts.68", "model.encoder.language_model.layers.6.experts.69", "model.encoder.language_model.layers.6.experts.70", "model.encoder.language_model.layers.6.experts.71", "model.encoder.language_model.layers.6.experts.72", "model.encoder.language_model.layers.6.experts.73", "model.encoder.language_model.layers.6.experts.74", "model.encoder.language_model.layers.6.experts.75", "model.encoder.language_model.layers.6.experts.76", "model.encoder.language_model.layers.6.experts.77", "model.encoder.language_model.layers.6.experts.78", "model.encoder.language_model.layers.6.experts.79", "model.encoder.language_model.layers.6.experts.80", "model.encoder.language_model.layers.6.experts.81", "model.encoder.language_model.layers.6.experts.82", "model.encoder.language_model.layers.6.experts.83", "model.encoder.language_model.layers.6.experts.84", "model.encoder.language_model.layers.6.experts.85", "model.encoder.language_model.layers.6.experts.86", "model.encoder.language_model.layers.6.experts.87", "model.encoder.language_model.layers.6.experts.88", "model.encoder.language_model.layers.6.experts.89", "model.encoder.language_model.layers.6.experts.90", "model.encoder.language_model.layers.6.experts.91", "model.encoder.language_model.layers.6.experts.92", "model.encoder.language_model.layers.6.experts.93", "model.encoder.language_model.layers.6.experts.94", "model.encoder.language_model.layers.6.experts.95", "model.encoder.language_model.layers.6.experts.96", "model.encoder.language_model.layers.6.experts.97", "model.encoder.language_model.layers.6.experts.98", "model.encoder.language_model.layers.6.experts.99", "model.encoder.language_model.layers.6.experts.100", "model.encoder.language_model.layers.6.experts.101", "model.encoder.language_model.layers.6.experts.102", "model.encoder.language_model.layers.6.experts.103", "model.encoder.language_model.layers.6.experts.104", "model.encoder.language_model.layers.6.experts.105", "model.encoder.language_model.layers.6.experts.106", "model.encoder.language_model.layers.6.experts.107", "model.encoder.language_model.layers.6.experts.108", "model.encoder.language_model.layers.6.experts.109", "model.encoder.language_model.layers.6.experts.110", "model.encoder.language_model.layers.6.experts.111", "model.encoder.language_model.layers.6.experts.112", "model.encoder.language_model.layers.6.experts.113", "model.encoder.language_model.layers.6.experts.114", "model.encoder.language_model.layers.6.experts.115", "model.encoder.language_model.layers.6.experts.116", "model.encoder.language_model.layers.6.experts.117", "model.encoder.language_model.layers.6.experts.118", "model.encoder.language_model.layers.6.experts.119", "model.encoder.language_model.layers.6.experts.120", "model.encoder.language_model.layers.6.experts.121", "model.encoder.language_model.layers.6.experts.122", "model.encoder.language_model.layers.6.experts.123", "model.encoder.language_model.layers.6.experts.124", "model.encoder.language_model.layers.6.experts.125", "model.encoder.language_model.layers.6.experts.126", "model.encoder.language_model.layers.6.experts.127", "model.encoder.language_model.layers.7.router", "model.encoder.language_model.layers.7.router.proj", "model.encoder.language_model.layers.7.experts.0", "model.encoder.language_model.layers.7.experts.1", "model.encoder.language_model.layers.7.experts.2", "model.encoder.language_model.layers.7.experts.3", "model.encoder.language_model.layers.7.experts.4", "model.encoder.language_model.layers.7.experts.5", "model.encoder.language_model.layers.7.experts.6", "model.encoder.language_model.layers.7.experts.7", "model.encoder.language_model.layers.7.experts.8", "model.encoder.language_model.layers.7.experts.9", "model.encoder.language_model.layers.7.experts.10", "model.encoder.language_model.layers.7.experts.11", "model.encoder.language_model.layers.7.experts.12", "model.encoder.language_model.layers.7.experts.13", "model.encoder.language_model.layers.7.experts.14", "model.encoder.language_model.layers.7.experts.15", "model.encoder.language_model.layers.7.experts.16", "model.encoder.language_model.layers.7.experts.17", "model.encoder.language_model.layers.7.experts.18", "model.encoder.language_model.layers.7.experts.19", "model.encoder.language_model.layers.7.experts.20", "model.encoder.language_model.layers.7.experts.21", "model.encoder.language_model.layers.7.experts.22", "model.encoder.language_model.layers.7.experts.23", "model.encoder.language_model.layers.7.experts.24", "model.encoder.language_model.layers.7.experts.25", "model.encoder.language_model.layers.7.experts.26", "model.encoder.language_model.layers.7.experts.27", "model.encoder.language_model.layers.7.experts.28", "model.encoder.language_model.layers.7.experts.29", "model.encoder.language_model.layers.7.experts.30", "model.encoder.language_model.layers.7.experts.31", "model.encoder.language_model.layers.7.experts.32", "model.encoder.language_model.layers.7.experts.33", "model.encoder.language_model.layers.7.experts.34", "model.encoder.language_model.layers.7.experts.35", "model.encoder.language_model.layers.7.experts.36", "model.encoder.language_model.layers.7.experts.37", "model.encoder.language_model.layers.7.experts.38", "model.encoder.language_model.layers.7.experts.39", "model.encoder.language_model.layers.7.experts.40", "model.encoder.language_model.layers.7.experts.41", "model.encoder.language_model.layers.7.experts.42", "model.encoder.language_model.layers.7.experts.43", "model.encoder.language_model.layers.7.experts.44", "model.encoder.language_model.layers.7.experts.45", "model.encoder.language_model.layers.7.experts.46", "model.encoder.language_model.layers.7.experts.47", "model.encoder.language_model.layers.7.experts.48", "model.encoder.language_model.layers.7.experts.49", "model.encoder.language_model.layers.7.experts.50", "model.encoder.language_model.layers.7.experts.51", "model.encoder.language_model.layers.7.experts.52", "model.encoder.language_model.layers.7.experts.53", "model.encoder.language_model.layers.7.experts.54", "model.encoder.language_model.layers.7.experts.55", "model.encoder.language_model.layers.7.experts.56", "model.encoder.language_model.layers.7.experts.57", "model.encoder.language_model.layers.7.experts.58", "model.encoder.language_model.layers.7.experts.59", "model.encoder.language_model.layers.7.experts.60", "model.encoder.language_model.layers.7.experts.61", "model.encoder.language_model.layers.7.experts.62", "model.encoder.language_model.layers.7.experts.63", "model.encoder.language_model.layers.7.experts.64", "model.encoder.language_model.layers.7.experts.65", "model.encoder.language_model.layers.7.experts.66", "model.encoder.language_model.layers.7.experts.67", "model.encoder.language_model.layers.7.experts.68", "model.encoder.language_model.layers.7.experts.69", "model.encoder.language_model.layers.7.experts.70", "model.encoder.language_model.layers.7.experts.71", "model.encoder.language_model.layers.7.experts.72", "model.encoder.language_model.layers.7.experts.73", "model.encoder.language_model.layers.7.experts.74", "model.encoder.language_model.layers.7.experts.75", "model.encoder.language_model.layers.7.experts.76", "model.encoder.language_model.layers.7.experts.77", "model.encoder.language_model.layers.7.experts.78", "model.encoder.language_model.layers.7.experts.79", "model.encoder.language_model.layers.7.experts.80", "model.encoder.language_model.layers.7.experts.81", "model.encoder.language_model.layers.7.experts.82", "model.encoder.language_model.layers.7.experts.83", "model.encoder.language_model.layers.7.experts.84", "model.encoder.language_model.layers.7.experts.85", "model.encoder.language_model.layers.7.experts.86", "model.encoder.language_model.layers.7.experts.87", "model.encoder.language_model.layers.7.experts.88", "model.encoder.language_model.layers.7.experts.89", "model.encoder.language_model.layers.7.experts.90", "model.encoder.language_model.layers.7.experts.91", "model.encoder.language_model.layers.7.experts.92", "model.encoder.language_model.layers.7.experts.93", "model.encoder.language_model.layers.7.experts.94", "model.encoder.language_model.layers.7.experts.95", "model.encoder.language_model.layers.7.experts.96", "model.encoder.language_model.layers.7.experts.97", "model.encoder.language_model.layers.7.experts.98", "model.encoder.language_model.layers.7.experts.99", "model.encoder.language_model.layers.7.experts.100", "model.encoder.language_model.layers.7.experts.101", "model.encoder.language_model.layers.7.experts.102", "model.encoder.language_model.layers.7.experts.103", "model.encoder.language_model.layers.7.experts.104", "model.encoder.language_model.layers.7.experts.105", "model.encoder.language_model.layers.7.experts.106", "model.encoder.language_model.layers.7.experts.107", "model.encoder.language_model.layers.7.experts.108", "model.encoder.language_model.layers.7.experts.109", "model.encoder.language_model.layers.7.experts.110", "model.encoder.language_model.layers.7.experts.111", "model.encoder.language_model.layers.7.experts.112", "model.encoder.language_model.layers.7.experts.113", "model.encoder.language_model.layers.7.experts.114", "model.encoder.language_model.layers.7.experts.115", "model.encoder.language_model.layers.7.experts.116", "model.encoder.language_model.layers.7.experts.117", "model.encoder.language_model.layers.7.experts.118", "model.encoder.language_model.layers.7.experts.119", "model.encoder.language_model.layers.7.experts.120", "model.encoder.language_model.layers.7.experts.121", "model.encoder.language_model.layers.7.experts.122", "model.encoder.language_model.layers.7.experts.123", "model.encoder.language_model.layers.7.experts.124", "model.encoder.language_model.layers.7.experts.125", "model.encoder.language_model.layers.7.experts.126", "model.encoder.language_model.layers.7.experts.127", "model.encoder.language_model.layers.8.router", "model.encoder.language_model.layers.8.router.proj", "model.encoder.language_model.layers.8.experts.0", "model.encoder.language_model.layers.8.experts.1", "model.encoder.language_model.layers.8.experts.2", "model.encoder.language_model.layers.8.experts.3", "model.encoder.language_model.layers.8.experts.4", "model.encoder.language_model.layers.8.experts.5", "model.encoder.language_model.layers.8.experts.6", "model.encoder.language_model.layers.8.experts.7", "model.encoder.language_model.layers.8.experts.8", "model.encoder.language_model.layers.8.experts.9", "model.encoder.language_model.layers.8.experts.10", "model.encoder.language_model.layers.8.experts.11", "model.encoder.language_model.layers.8.experts.12", "model.encoder.language_model.layers.8.experts.13", "model.encoder.language_model.layers.8.experts.14", "model.encoder.language_model.layers.8.experts.15", "model.encoder.language_model.layers.8.experts.16", "model.encoder.language_model.layers.8.experts.17", "model.encoder.language_model.layers.8.experts.18", "model.encoder.language_model.layers.8.experts.19", "model.encoder.language_model.layers.8.experts.20", "model.encoder.language_model.layers.8.experts.21", "model.encoder.language_model.layers.8.experts.22", "model.encoder.language_model.layers.8.experts.23", "model.encoder.language_model.layers.8.experts.24", "model.encoder.language_model.layers.8.experts.25", "model.encoder.language_model.layers.8.experts.26", "model.encoder.language_model.layers.8.experts.27", "model.encoder.language_model.layers.8.experts.28", "model.encoder.language_model.layers.8.experts.29", "model.encoder.language_model.layers.8.experts.30", "model.encoder.language_model.layers.8.experts.31", "model.encoder.language_model.layers.8.experts.32", "model.encoder.language_model.layers.8.experts.33", "model.encoder.language_model.layers.8.experts.34", "model.encoder.language_model.layers.8.experts.35", "model.encoder.language_model.layers.8.experts.36", "model.encoder.language_model.layers.8.experts.37", "model.encoder.language_model.layers.8.experts.38", "model.encoder.language_model.layers.8.experts.39", "model.encoder.language_model.layers.8.experts.40", "model.encoder.language_model.layers.8.experts.41", "model.encoder.language_model.layers.8.experts.42", "model.encoder.language_model.layers.8.experts.43", "model.encoder.language_model.layers.8.experts.44", "model.encoder.language_model.layers.8.experts.45", "model.encoder.language_model.layers.8.experts.46", "model.encoder.language_model.layers.8.experts.47", "model.encoder.language_model.layers.8.experts.48", "model.encoder.language_model.layers.8.experts.49", "model.encoder.language_model.layers.8.experts.50", "model.encoder.language_model.layers.8.experts.51", "model.encoder.language_model.layers.8.experts.52", "model.encoder.language_model.layers.8.experts.53", "model.encoder.language_model.layers.8.experts.54", "model.encoder.language_model.layers.8.experts.55", "model.encoder.language_model.layers.8.experts.56", "model.encoder.language_model.layers.8.experts.57", "model.encoder.language_model.layers.8.experts.58", "model.encoder.language_model.layers.8.experts.59", "model.encoder.language_model.layers.8.experts.60", "model.encoder.language_model.layers.8.experts.61", "model.encoder.language_model.layers.8.experts.62", "model.encoder.language_model.layers.8.experts.63", "model.encoder.language_model.layers.8.experts.64", "model.encoder.language_model.layers.8.experts.65", "model.encoder.language_model.layers.8.experts.66", "model.encoder.language_model.layers.8.experts.67", "model.encoder.language_model.layers.8.experts.68", "model.encoder.language_model.layers.8.experts.69", "model.encoder.language_model.layers.8.experts.70", "model.encoder.language_model.layers.8.experts.71", "model.encoder.language_model.layers.8.experts.72", "model.encoder.language_model.layers.8.experts.73", "model.encoder.language_model.layers.8.experts.74", "model.encoder.language_model.layers.8.experts.75", "model.encoder.language_model.layers.8.experts.76", "model.encoder.language_model.layers.8.experts.77", "model.encoder.language_model.layers.8.experts.78", "model.encoder.language_model.layers.8.experts.79", "model.encoder.language_model.layers.8.experts.80", "model.encoder.language_model.layers.8.experts.81", "model.encoder.language_model.layers.8.experts.82", "model.encoder.language_model.layers.8.experts.83", "model.encoder.language_model.layers.8.experts.84", "model.encoder.language_model.layers.8.experts.85", "model.encoder.language_model.layers.8.experts.86", "model.encoder.language_model.layers.8.experts.87", "model.encoder.language_model.layers.8.experts.88", "model.encoder.language_model.layers.8.experts.89", "model.encoder.language_model.layers.8.experts.90", "model.encoder.language_model.layers.8.experts.91", "model.encoder.language_model.layers.8.experts.92", "model.encoder.language_model.layers.8.experts.93", "model.encoder.language_model.layers.8.experts.94", "model.encoder.language_model.layers.8.experts.95", "model.encoder.language_model.layers.8.experts.96", "model.encoder.language_model.layers.8.experts.97", "model.encoder.language_model.layers.8.experts.98", "model.encoder.language_model.layers.8.experts.99", "model.encoder.language_model.layers.8.experts.100", "model.encoder.language_model.layers.8.experts.101", "model.encoder.language_model.layers.8.experts.102", "model.encoder.language_model.layers.8.experts.103", "model.encoder.language_model.layers.8.experts.104", "model.encoder.language_model.layers.8.experts.105", "model.encoder.language_model.layers.8.experts.106", "model.encoder.language_model.layers.8.experts.107", "model.encoder.language_model.layers.8.experts.108", "model.encoder.language_model.layers.8.experts.109", "model.encoder.language_model.layers.8.experts.110", "model.encoder.language_model.layers.8.experts.111", "model.encoder.language_model.layers.8.experts.112", "model.encoder.language_model.layers.8.experts.113", "model.encoder.language_model.layers.8.experts.114", "model.encoder.language_model.layers.8.experts.115", "model.encoder.language_model.layers.8.experts.116", "model.encoder.language_model.layers.8.experts.117", "model.encoder.language_model.layers.8.experts.118", "model.encoder.language_model.layers.8.experts.119", "model.encoder.language_model.layers.8.experts.120", "model.encoder.language_model.layers.8.experts.121", "model.encoder.language_model.layers.8.experts.122", "model.encoder.language_model.layers.8.experts.123", "model.encoder.language_model.layers.8.experts.124", "model.encoder.language_model.layers.8.experts.125", "model.encoder.language_model.layers.8.experts.126", "model.encoder.language_model.layers.8.experts.127", "model.encoder.language_model.layers.9.router", "model.encoder.language_model.layers.9.router.proj", "model.encoder.language_model.layers.9.experts.0", "model.encoder.language_model.layers.9.experts.1", "model.encoder.language_model.layers.9.experts.2", "model.encoder.language_model.layers.9.experts.3", "model.encoder.language_model.layers.9.experts.4", "model.encoder.language_model.layers.9.experts.5", "model.encoder.language_model.layers.9.experts.6", "model.encoder.language_model.layers.9.experts.7", "model.encoder.language_model.layers.9.experts.8", "model.encoder.language_model.layers.9.experts.9", "model.encoder.language_model.layers.9.experts.10", "model.encoder.language_model.layers.9.experts.11", "model.encoder.language_model.layers.9.experts.12", "model.encoder.language_model.layers.9.experts.13", "model.encoder.language_model.layers.9.experts.14", "model.encoder.language_model.layers.9.experts.15", "model.encoder.language_model.layers.9.experts.16", "model.encoder.language_model.layers.9.experts.17", "model.encoder.language_model.layers.9.experts.18", "model.encoder.language_model.layers.9.experts.19", "model.encoder.language_model.layers.9.experts.20", "model.encoder.language_model.layers.9.experts.21", "model.encoder.language_model.layers.9.experts.22", "model.encoder.language_model.layers.9.experts.23", "model.encoder.language_model.layers.9.experts.24", "model.encoder.language_model.layers.9.experts.25", "model.encoder.language_model.layers.9.experts.26", "model.encoder.language_model.layers.9.experts.27", "model.encoder.language_model.layers.9.experts.28", "model.encoder.language_model.layers.9.experts.29", "model.encoder.language_model.layers.9.experts.30", "model.encoder.language_model.layers.9.experts.31", "model.encoder.language_model.layers.9.experts.32", "model.encoder.language_model.layers.9.experts.33", "model.encoder.language_model.layers.9.experts.34", "model.encoder.language_model.layers.9.experts.35", "model.encoder.language_model.layers.9.experts.36", "model.encoder.language_model.layers.9.experts.37", "model.encoder.language_model.layers.9.experts.38", "model.encoder.language_model.layers.9.experts.39", "model.encoder.language_model.layers.9.experts.40", "model.encoder.language_model.layers.9.experts.41", "model.encoder.language_model.layers.9.experts.42", "model.encoder.language_model.layers.9.experts.43", "model.encoder.language_model.layers.9.experts.44", "model.encoder.language_model.layers.9.experts.45", "model.encoder.language_model.layers.9.experts.46", "model.encoder.language_model.layers.9.experts.47", "model.encoder.language_model.layers.9.experts.48", "model.encoder.language_model.layers.9.experts.49", "model.encoder.language_model.layers.9.experts.50", "model.encoder.language_model.layers.9.experts.51", "model.encoder.language_model.layers.9.experts.52", "model.encoder.language_model.layers.9.experts.53", "model.encoder.language_model.layers.9.experts.54", "model.encoder.language_model.layers.9.experts.55", "model.encoder.language_model.layers.9.experts.56", "model.encoder.language_model.layers.9.experts.57", "model.encoder.language_model.layers.9.experts.58", "model.encoder.language_model.layers.9.experts.59", "model.encoder.language_model.layers.9.experts.60", "model.encoder.language_model.layers.9.experts.61", "model.encoder.language_model.layers.9.experts.62", "model.encoder.language_model.layers.9.experts.63", "model.encoder.language_model.layers.9.experts.64", "model.encoder.language_model.layers.9.experts.65", "model.encoder.language_model.layers.9.experts.66", "model.encoder.language_model.layers.9.experts.67", "model.encoder.language_model.layers.9.experts.68", "model.encoder.language_model.layers.9.experts.69", "model.encoder.language_model.layers.9.experts.70", "model.encoder.language_model.layers.9.experts.71", "model.encoder.language_model.layers.9.experts.72", "model.encoder.language_model.layers.9.experts.73", "model.encoder.language_model.layers.9.experts.74", "model.encoder.language_model.layers.9.experts.75", "model.encoder.language_model.layers.9.experts.76", "model.encoder.language_model.layers.9.experts.77", "model.encoder.language_model.layers.9.experts.78", "model.encoder.language_model.layers.9.experts.79", "model.encoder.language_model.layers.9.experts.80", "model.encoder.language_model.layers.9.experts.81", "model.encoder.language_model.layers.9.experts.82", "model.encoder.language_model.layers.9.experts.83", "model.encoder.language_model.layers.9.experts.84", "model.encoder.language_model.layers.9.experts.85", "model.encoder.language_model.layers.9.experts.86", "model.encoder.language_model.layers.9.experts.87", "model.encoder.language_model.layers.9.experts.88", "model.encoder.language_model.layers.9.experts.89", "model.encoder.language_model.layers.9.experts.90", "model.encoder.language_model.layers.9.experts.91", "model.encoder.language_model.layers.9.experts.92", "model.encoder.language_model.layers.9.experts.93", "model.encoder.language_model.layers.9.experts.94", "model.encoder.language_model.layers.9.experts.95", "model.encoder.language_model.layers.9.experts.96", "model.encoder.language_model.layers.9.experts.97", "model.encoder.language_model.layers.9.experts.98", "model.encoder.language_model.layers.9.experts.99", "model.encoder.language_model.layers.9.experts.100", "model.encoder.language_model.layers.9.experts.101", "model.encoder.language_model.layers.9.experts.102", "model.encoder.language_model.layers.9.experts.103", "model.encoder.language_model.layers.9.experts.104", "model.encoder.language_model.layers.9.experts.105", "model.encoder.language_model.layers.9.experts.106", "model.encoder.language_model.layers.9.experts.107", "model.encoder.language_model.layers.9.experts.108", "model.encoder.language_model.layers.9.experts.109", "model.encoder.language_model.layers.9.experts.110", "model.encoder.language_model.layers.9.experts.111", "model.encoder.language_model.layers.9.experts.112", "model.encoder.language_model.layers.9.experts.113", "model.encoder.language_model.layers.9.experts.114", "model.encoder.language_model.layers.9.experts.115", "model.encoder.language_model.layers.9.experts.116", "model.encoder.language_model.layers.9.experts.117", "model.encoder.language_model.layers.9.experts.118", "model.encoder.language_model.layers.9.experts.119", "model.encoder.language_model.layers.9.experts.120", "model.encoder.language_model.layers.9.experts.121", "model.encoder.language_model.layers.9.experts.122", "model.encoder.language_model.layers.9.experts.123", "model.encoder.language_model.layers.9.experts.124", "model.encoder.language_model.layers.9.experts.125", "model.encoder.language_model.layers.9.experts.126", "model.encoder.language_model.layers.9.experts.127", "model.encoder.language_model.layers.10.router", "model.encoder.language_model.layers.10.router.proj", "model.encoder.language_model.layers.10.experts.0", "model.encoder.language_model.layers.10.experts.1", "model.encoder.language_model.layers.10.experts.2", "model.encoder.language_model.layers.10.experts.3", "model.encoder.language_model.layers.10.experts.4", "model.encoder.language_model.layers.10.experts.5", "model.encoder.language_model.layers.10.experts.6", "model.encoder.language_model.layers.10.experts.7", "model.encoder.language_model.layers.10.experts.8", "model.encoder.language_model.layers.10.experts.9", "model.encoder.language_model.layers.10.experts.10", "model.encoder.language_model.layers.10.experts.11", "model.encoder.language_model.layers.10.experts.12", "model.encoder.language_model.layers.10.experts.13", "model.encoder.language_model.layers.10.experts.14", "model.encoder.language_model.layers.10.experts.15", "model.encoder.language_model.layers.10.experts.16", "model.encoder.language_model.layers.10.experts.17", "model.encoder.language_model.layers.10.experts.18", "model.encoder.language_model.layers.10.experts.19", "model.encoder.language_model.layers.10.experts.20", "model.encoder.language_model.layers.10.experts.21", "model.encoder.language_model.layers.10.experts.22", "model.encoder.language_model.layers.10.experts.23", "model.encoder.language_model.layers.10.experts.24", "model.encoder.language_model.layers.10.experts.25", "model.encoder.language_model.layers.10.experts.26", "model.encoder.language_model.layers.10.experts.27", "model.encoder.language_model.layers.10.experts.28", "model.encoder.language_model.layers.10.experts.29", "model.encoder.language_model.layers.10.experts.30", "model.encoder.language_model.layers.10.experts.31", "model.encoder.language_model.layers.10.experts.32", "model.encoder.language_model.layers.10.experts.33", "model.encoder.language_model.layers.10.experts.34", "model.encoder.language_model.layers.10.experts.35", "model.encoder.language_model.layers.10.experts.36", "model.encoder.language_model.layers.10.experts.37", "model.encoder.language_model.layers.10.experts.38", "model.encoder.language_model.layers.10.experts.39", "model.encoder.language_model.layers.10.experts.40", "model.encoder.language_model.layers.10.experts.41", "model.encoder.language_model.layers.10.experts.42", "model.encoder.language_model.layers.10.experts.43", "model.encoder.language_model.layers.10.experts.44", "model.encoder.language_model.layers.10.experts.45", "model.encoder.language_model.layers.10.experts.46", "model.encoder.language_model.layers.10.experts.47", "model.encoder.language_model.layers.10.experts.48", "model.encoder.language_model.layers.10.experts.49", "model.encoder.language_model.layers.10.experts.50", "model.encoder.language_model.layers.10.experts.51", "model.encoder.language_model.layers.10.experts.52", "model.encoder.language_model.layers.10.experts.53", "model.encoder.language_model.layers.10.experts.54", "model.encoder.language_model.layers.10.experts.55", "model.encoder.language_model.layers.10.experts.56", "model.encoder.language_model.layers.10.experts.57", "model.encoder.language_model.layers.10.experts.58", "model.encoder.language_model.layers.10.experts.59", "model.encoder.language_model.layers.10.experts.60", "model.encoder.language_model.layers.10.experts.61", "model.encoder.language_model.layers.10.experts.62", "model.encoder.language_model.layers.10.experts.63", "model.encoder.language_model.layers.10.experts.64", "model.encoder.language_model.layers.10.experts.65", "model.encoder.language_model.layers.10.experts.66", "model.encoder.language_model.layers.10.experts.67", "model.encoder.language_model.layers.10.experts.68", "model.encoder.language_model.layers.10.experts.69", "model.encoder.language_model.layers.10.experts.70", "model.encoder.language_model.layers.10.experts.71", "model.encoder.language_model.layers.10.experts.72", "model.encoder.language_model.layers.10.experts.73", "model.encoder.language_model.layers.10.experts.74", "model.encoder.language_model.layers.10.experts.75", "model.encoder.language_model.layers.10.experts.76", "model.encoder.language_model.layers.10.experts.77", "model.encoder.language_model.layers.10.experts.78", "model.encoder.language_model.layers.10.experts.79", "model.encoder.language_model.layers.10.experts.80", "model.encoder.language_model.layers.10.experts.81", "model.encoder.language_model.layers.10.experts.82", "model.encoder.language_model.layers.10.experts.83", "model.encoder.language_model.layers.10.experts.84", "model.encoder.language_model.layers.10.experts.85", "model.encoder.language_model.layers.10.experts.86", "model.encoder.language_model.layers.10.experts.87", "model.encoder.language_model.layers.10.experts.88", "model.encoder.language_model.layers.10.experts.89", "model.encoder.language_model.layers.10.experts.90", "model.encoder.language_model.layers.10.experts.91", "model.encoder.language_model.layers.10.experts.92", "model.encoder.language_model.layers.10.experts.93", "model.encoder.language_model.layers.10.experts.94", "model.encoder.language_model.layers.10.experts.95", "model.encoder.language_model.layers.10.experts.96", "model.encoder.language_model.layers.10.experts.97", "model.encoder.language_model.layers.10.experts.98", "model.encoder.language_model.layers.10.experts.99", "model.encoder.language_model.layers.10.experts.100", "model.encoder.language_model.layers.10.experts.101", "model.encoder.language_model.layers.10.experts.102", "model.encoder.language_model.layers.10.experts.103", "model.encoder.language_model.layers.10.experts.104", "model.encoder.language_model.layers.10.experts.105", "model.encoder.language_model.layers.10.experts.106", "model.encoder.language_model.layers.10.experts.107", "model.encoder.language_model.layers.10.experts.108", "model.encoder.language_model.layers.10.experts.109", "model.encoder.language_model.layers.10.experts.110", "model.encoder.language_model.layers.10.experts.111", "model.encoder.language_model.layers.10.experts.112", "model.encoder.language_model.layers.10.experts.113", "model.encoder.language_model.layers.10.experts.114", "model.encoder.language_model.layers.10.experts.115", "model.encoder.language_model.layers.10.experts.116", "model.encoder.language_model.layers.10.experts.117", "model.encoder.language_model.layers.10.experts.118", "model.encoder.language_model.layers.10.experts.119", "model.encoder.language_model.layers.10.experts.120", "model.encoder.language_model.layers.10.experts.121", "model.encoder.language_model.layers.10.experts.122", "model.encoder.language_model.layers.10.experts.123", "model.encoder.language_model.layers.10.experts.124", "model.encoder.language_model.layers.10.experts.125", "model.encoder.language_model.layers.10.experts.126", "model.encoder.language_model.layers.10.experts.127", "model.encoder.language_model.layers.11.router", "model.encoder.language_model.layers.11.router.proj", "model.encoder.language_model.layers.11.experts.0", "model.encoder.language_model.layers.11.experts.1", "model.encoder.language_model.layers.11.experts.2", "model.encoder.language_model.layers.11.experts.3", "model.encoder.language_model.layers.11.experts.4", "model.encoder.language_model.layers.11.experts.5", "model.encoder.language_model.layers.11.experts.6", "model.encoder.language_model.layers.11.experts.7", "model.encoder.language_model.layers.11.experts.8", "model.encoder.language_model.layers.11.experts.9", "model.encoder.language_model.layers.11.experts.10", "model.encoder.language_model.layers.11.experts.11", "model.encoder.language_model.layers.11.experts.12", "model.encoder.language_model.layers.11.experts.13", "model.encoder.language_model.layers.11.experts.14", "model.encoder.language_model.layers.11.experts.15", "model.encoder.language_model.layers.11.experts.16", "model.encoder.language_model.layers.11.experts.17", "model.encoder.language_model.layers.11.experts.18", "model.encoder.language_model.layers.11.experts.19", "model.encoder.language_model.layers.11.experts.20", "model.encoder.language_model.layers.11.experts.21", "model.encoder.language_model.layers.11.experts.22", "model.encoder.language_model.layers.11.experts.23", "model.encoder.language_model.layers.11.experts.24", "model.encoder.language_model.layers.11.experts.25", "model.encoder.language_model.layers.11.experts.26", "model.encoder.language_model.layers.11.experts.27", "model.encoder.language_model.layers.11.experts.28", "model.encoder.language_model.layers.11.experts.29", "model.encoder.language_model.layers.11.experts.30", "model.encoder.language_model.layers.11.experts.31", "model.encoder.language_model.layers.11.experts.32", "model.encoder.language_model.layers.11.experts.33", "model.encoder.language_model.layers.11.experts.34", "model.encoder.language_model.layers.11.experts.35", "model.encoder.language_model.layers.11.experts.36", "model.encoder.language_model.layers.11.experts.37", "model.encoder.language_model.layers.11.experts.38", "model.encoder.language_model.layers.11.experts.39", "model.encoder.language_model.layers.11.experts.40", "model.encoder.language_model.layers.11.experts.41", "model.encoder.language_model.layers.11.experts.42", "model.encoder.language_model.layers.11.experts.43", "model.encoder.language_model.layers.11.experts.44", "model.encoder.language_model.layers.11.experts.45", "model.encoder.language_model.layers.11.experts.46", "model.encoder.language_model.layers.11.experts.47", "model.encoder.language_model.layers.11.experts.48", "model.encoder.language_model.layers.11.experts.49", "model.encoder.language_model.layers.11.experts.50", "model.encoder.language_model.layers.11.experts.51", "model.encoder.language_model.layers.11.experts.52", "model.encoder.language_model.layers.11.experts.53", "model.encoder.language_model.layers.11.experts.54", "model.encoder.language_model.layers.11.experts.55", "model.encoder.language_model.layers.11.experts.56", "model.encoder.language_model.layers.11.experts.57", "model.encoder.language_model.layers.11.experts.58", "model.encoder.language_model.layers.11.experts.59", "model.encoder.language_model.layers.11.experts.60", "model.encoder.language_model.layers.11.experts.61", "model.encoder.language_model.layers.11.experts.62", "model.encoder.language_model.layers.11.experts.63", "model.encoder.language_model.layers.11.experts.64", "model.encoder.language_model.layers.11.experts.65", "model.encoder.language_model.layers.11.experts.66", "model.encoder.language_model.layers.11.experts.67", "model.encoder.language_model.layers.11.experts.68", "model.encoder.language_model.layers.11.experts.69", "model.encoder.language_model.layers.11.experts.70", "model.encoder.language_model.layers.11.experts.71", "model.encoder.language_model.layers.11.experts.72", "model.encoder.language_model.layers.11.experts.73", "model.encoder.language_model.layers.11.experts.74", "model.encoder.language_model.layers.11.experts.75", "model.encoder.language_model.layers.11.experts.76", "model.encoder.language_model.layers.11.experts.77", "model.encoder.language_model.layers.11.experts.78", "model.encoder.language_model.layers.11.experts.79", "model.encoder.language_model.layers.11.experts.80", "model.encoder.language_model.layers.11.experts.81", "model.encoder.language_model.layers.11.experts.82", "model.encoder.language_model.layers.11.experts.83", "model.encoder.language_model.layers.11.experts.84", "model.encoder.language_model.layers.11.experts.85", "model.encoder.language_model.layers.11.experts.86", "model.encoder.language_model.layers.11.experts.87", "model.encoder.language_model.layers.11.experts.88", "model.encoder.language_model.layers.11.experts.89", "model.encoder.language_model.layers.11.experts.90", "model.encoder.language_model.layers.11.experts.91", "model.encoder.language_model.layers.11.experts.92", "model.encoder.language_model.layers.11.experts.93", "model.encoder.language_model.layers.11.experts.94", "model.encoder.language_model.layers.11.experts.95", "model.encoder.language_model.layers.11.experts.96", "model.encoder.language_model.layers.11.experts.97", "model.encoder.language_model.layers.11.experts.98", "model.encoder.language_model.layers.11.experts.99", "model.encoder.language_model.layers.11.experts.100", "model.encoder.language_model.layers.11.experts.101", "model.encoder.language_model.layers.11.experts.102", "model.encoder.language_model.layers.11.experts.103", "model.encoder.language_model.layers.11.experts.104", "model.encoder.language_model.layers.11.experts.105", "model.encoder.language_model.layers.11.experts.106", "model.encoder.language_model.layers.11.experts.107", "model.encoder.language_model.layers.11.experts.108", "model.encoder.language_model.layers.11.experts.109", "model.encoder.language_model.layers.11.experts.110", "model.encoder.language_model.layers.11.experts.111", "model.encoder.language_model.layers.11.experts.112", "model.encoder.language_model.layers.11.experts.113", "model.encoder.language_model.layers.11.experts.114", "model.encoder.language_model.layers.11.experts.115", "model.encoder.language_model.layers.11.experts.116", "model.encoder.language_model.layers.11.experts.117", "model.encoder.language_model.layers.11.experts.118", "model.encoder.language_model.layers.11.experts.119", "model.encoder.language_model.layers.11.experts.120", "model.encoder.language_model.layers.11.experts.121", "model.encoder.language_model.layers.11.experts.122", "model.encoder.language_model.layers.11.experts.123", "model.encoder.language_model.layers.11.experts.124", "model.encoder.language_model.layers.11.experts.125", "model.encoder.language_model.layers.11.experts.126", "model.encoder.language_model.layers.11.experts.127", "model.encoder.language_model.layers.12.router", "model.encoder.language_model.layers.12.router.proj", "model.encoder.language_model.layers.12.experts.0", "model.encoder.language_model.layers.12.experts.1", "model.encoder.language_model.layers.12.experts.2", "model.encoder.language_model.layers.12.experts.3", "model.encoder.language_model.layers.12.experts.4", "model.encoder.language_model.layers.12.experts.5", "model.encoder.language_model.layers.12.experts.6", "model.encoder.language_model.layers.12.experts.7", "model.encoder.language_model.layers.12.experts.8", "model.encoder.language_model.layers.12.experts.9", "model.encoder.language_model.layers.12.experts.10", "model.encoder.language_model.layers.12.experts.11", "model.encoder.language_model.layers.12.experts.12", "model.encoder.language_model.layers.12.experts.13", "model.encoder.language_model.layers.12.experts.14", "model.encoder.language_model.layers.12.experts.15", "model.encoder.language_model.layers.12.experts.16", "model.encoder.language_model.layers.12.experts.17", "model.encoder.language_model.layers.12.experts.18", "model.encoder.language_model.layers.12.experts.19", "model.encoder.language_model.layers.12.experts.20", "model.encoder.language_model.layers.12.experts.21", "model.encoder.language_model.layers.12.experts.22", "model.encoder.language_model.layers.12.experts.23", "model.encoder.language_model.layers.12.experts.24", "model.encoder.language_model.layers.12.experts.25", "model.encoder.language_model.layers.12.experts.26", "model.encoder.language_model.layers.12.experts.27", "model.encoder.language_model.layers.12.experts.28", "model.encoder.language_model.layers.12.experts.29", "model.encoder.language_model.layers.12.experts.30", "model.encoder.language_model.layers.12.experts.31", "model.encoder.language_model.layers.12.experts.32", "model.encoder.language_model.layers.12.experts.33", "model.encoder.language_model.layers.12.experts.34", "model.encoder.language_model.layers.12.experts.35", "model.encoder.language_model.layers.12.experts.36", "model.encoder.language_model.layers.12.experts.37", "model.encoder.language_model.layers.12.experts.38", "model.encoder.language_model.layers.12.experts.39", "model.encoder.language_model.layers.12.experts.40", "model.encoder.language_model.layers.12.experts.41", "model.encoder.language_model.layers.12.experts.42", "model.encoder.language_model.layers.12.experts.43", "model.encoder.language_model.layers.12.experts.44", "model.encoder.language_model.layers.12.experts.45", "model.encoder.language_model.layers.12.experts.46", "model.encoder.language_model.layers.12.experts.47", "model.encoder.language_model.layers.12.experts.48", "model.encoder.language_model.layers.12.experts.49", "model.encoder.language_model.layers.12.experts.50", "model.encoder.language_model.layers.12.experts.51", "model.encoder.language_model.layers.12.experts.52", "model.encoder.language_model.layers.12.experts.53", "model.encoder.language_model.layers.12.experts.54", "model.encoder.language_model.layers.12.experts.55", "model.encoder.language_model.layers.12.experts.56", "model.encoder.language_model.layers.12.experts.57", "model.encoder.language_model.layers.12.experts.58", "model.encoder.language_model.layers.12.experts.59", "model.encoder.language_model.layers.12.experts.60", "model.encoder.language_model.layers.12.experts.61", "model.encoder.language_model.layers.12.experts.62", "model.encoder.language_model.layers.12.experts.63", "model.encoder.language_model.layers.12.experts.64", "model.encoder.language_model.layers.12.experts.65", "model.encoder.language_model.layers.12.experts.66", "model.encoder.language_model.layers.12.experts.67", "model.encoder.language_model.layers.12.experts.68", "model.encoder.language_model.layers.12.experts.69", "model.encoder.language_model.layers.12.experts.70", "model.encoder.language_model.layers.12.experts.71", "model.encoder.language_model.layers.12.experts.72", "model.encoder.language_model.layers.12.experts.73", "model.encoder.language_model.layers.12.experts.74", "model.encoder.language_model.layers.12.experts.75", "model.encoder.language_model.layers.12.experts.76", "model.encoder.language_model.layers.12.experts.77", "model.encoder.language_model.layers.12.experts.78", "model.encoder.language_model.layers.12.experts.79", "model.encoder.language_model.layers.12.experts.80", "model.encoder.language_model.layers.12.experts.81", "model.encoder.language_model.layers.12.experts.82", "model.encoder.language_model.layers.12.experts.83", "model.encoder.language_model.layers.12.experts.84", "model.encoder.language_model.layers.12.experts.85", "model.encoder.language_model.layers.12.experts.86", "model.encoder.language_model.layers.12.experts.87", "model.encoder.language_model.layers.12.experts.88", "model.encoder.language_model.layers.12.experts.89", "model.encoder.language_model.layers.12.experts.90", "model.encoder.language_model.layers.12.experts.91", "model.encoder.language_model.layers.12.experts.92", "model.encoder.language_model.layers.12.experts.93", "model.encoder.language_model.layers.12.experts.94", "model.encoder.language_model.layers.12.experts.95", "model.encoder.language_model.layers.12.experts.96", "model.encoder.language_model.layers.12.experts.97", "model.encoder.language_model.layers.12.experts.98", "model.encoder.language_model.layers.12.experts.99", "model.encoder.language_model.layers.12.experts.100", "model.encoder.language_model.layers.12.experts.101", "model.encoder.language_model.layers.12.experts.102", "model.encoder.language_model.layers.12.experts.103", "model.encoder.language_model.layers.12.experts.104", "model.encoder.language_model.layers.12.experts.105", "model.encoder.language_model.layers.12.experts.106", "model.encoder.language_model.layers.12.experts.107", "model.encoder.language_model.layers.12.experts.108", "model.encoder.language_model.layers.12.experts.109", "model.encoder.language_model.layers.12.experts.110", "model.encoder.language_model.layers.12.experts.111", "model.encoder.language_model.layers.12.experts.112", "model.encoder.language_model.layers.12.experts.113", "model.encoder.language_model.layers.12.experts.114", "model.encoder.language_model.layers.12.experts.115", "model.encoder.language_model.layers.12.experts.116", "model.encoder.language_model.layers.12.experts.117", "model.encoder.language_model.layers.12.experts.118", "model.encoder.language_model.layers.12.experts.119", "model.encoder.language_model.layers.12.experts.120", "model.encoder.language_model.layers.12.experts.121", "model.encoder.language_model.layers.12.experts.122", "model.encoder.language_model.layers.12.experts.123", "model.encoder.language_model.layers.12.experts.124", "model.encoder.language_model.layers.12.experts.125", "model.encoder.language_model.layers.12.experts.126", "model.encoder.language_model.layers.12.experts.127", "model.encoder.language_model.layers.13.router", "model.encoder.language_model.layers.13.router.proj", "model.encoder.language_model.layers.13.experts.0", "model.encoder.language_model.layers.13.experts.1", "model.encoder.language_model.layers.13.experts.2", "model.encoder.language_model.layers.13.experts.3", "model.encoder.language_model.layers.13.experts.4", "model.encoder.language_model.layers.13.experts.5", "model.encoder.language_model.layers.13.experts.6", "model.encoder.language_model.layers.13.experts.7", "model.encoder.language_model.layers.13.experts.8", "model.encoder.language_model.layers.13.experts.9", "model.encoder.language_model.layers.13.experts.10", "model.encoder.language_model.layers.13.experts.11", "model.encoder.language_model.layers.13.experts.12", "model.encoder.language_model.layers.13.experts.13", "model.encoder.language_model.layers.13.experts.14", "model.encoder.language_model.layers.13.experts.15", "model.encoder.language_model.layers.13.experts.16", "model.encoder.language_model.layers.13.experts.17", "model.encoder.language_model.layers.13.experts.18", "model.encoder.language_model.layers.13.experts.19", "model.encoder.language_model.layers.13.experts.20", "model.encoder.language_model.layers.13.experts.21", "model.encoder.language_model.layers.13.experts.22", "model.encoder.language_model.layers.13.experts.23", "model.encoder.language_model.layers.13.experts.24", "model.encoder.language_model.layers.13.experts.25", "model.encoder.language_model.layers.13.experts.26", "model.encoder.language_model.layers.13.experts.27", "model.encoder.language_model.layers.13.experts.28", "model.encoder.language_model.layers.13.experts.29", "model.encoder.language_model.layers.13.experts.30", "model.encoder.language_model.layers.13.experts.31", "model.encoder.language_model.layers.13.experts.32", "model.encoder.language_model.layers.13.experts.33", "model.encoder.language_model.layers.13.experts.34", "model.encoder.language_model.layers.13.experts.35", "model.encoder.language_model.layers.13.experts.36", "model.encoder.language_model.layers.13.experts.37", "model.encoder.language_model.layers.13.experts.38", "model.encoder.language_model.layers.13.experts.39", "model.encoder.language_model.layers.13.experts.40", "model.encoder.language_model.layers.13.experts.41", "model.encoder.language_model.layers.13.experts.42", "model.encoder.language_model.layers.13.experts.43", "model.encoder.language_model.layers.13.experts.44", "model.encoder.language_model.layers.13.experts.45", "model.encoder.language_model.layers.13.experts.46", "model.encoder.language_model.layers.13.experts.47", "model.encoder.language_model.layers.13.experts.48", "model.encoder.language_model.layers.13.experts.49", "model.encoder.language_model.layers.13.experts.50", "model.encoder.language_model.layers.13.experts.51", "model.encoder.language_model.layers.13.experts.52", "model.encoder.language_model.layers.13.experts.53", "model.encoder.language_model.layers.13.experts.54", "model.encoder.language_model.layers.13.experts.55", "model.encoder.language_model.layers.13.experts.56", "model.encoder.language_model.layers.13.experts.57", "model.encoder.language_model.layers.13.experts.58", "model.encoder.language_model.layers.13.experts.59", "model.encoder.language_model.layers.13.experts.60", "model.encoder.language_model.layers.13.experts.61", "model.encoder.language_model.layers.13.experts.62", "model.encoder.language_model.layers.13.experts.63", "model.encoder.language_model.layers.13.experts.64", "model.encoder.language_model.layers.13.experts.65", "model.encoder.language_model.layers.13.experts.66", "model.encoder.language_model.layers.13.experts.67", "model.encoder.language_model.layers.13.experts.68", "model.encoder.language_model.layers.13.experts.69", "model.encoder.language_model.layers.13.experts.70", "model.encoder.language_model.layers.13.experts.71", "model.encoder.language_model.layers.13.experts.72", "model.encoder.language_model.layers.13.experts.73", "model.encoder.language_model.layers.13.experts.74", "model.encoder.language_model.layers.13.experts.75", "model.encoder.language_model.layers.13.experts.76", "model.encoder.language_model.layers.13.experts.77", "model.encoder.language_model.layers.13.experts.78", "model.encoder.language_model.layers.13.experts.79", "model.encoder.language_model.layers.13.experts.80", "model.encoder.language_model.layers.13.experts.81", "model.encoder.language_model.layers.13.experts.82", "model.encoder.language_model.layers.13.experts.83", "model.encoder.language_model.layers.13.experts.84", "model.encoder.language_model.layers.13.experts.85", "model.encoder.language_model.layers.13.experts.86", "model.encoder.language_model.layers.13.experts.87", "model.encoder.language_model.layers.13.experts.88", "model.encoder.language_model.layers.13.experts.89", "model.encoder.language_model.layers.13.experts.90", "model.encoder.language_model.layers.13.experts.91", "model.encoder.language_model.layers.13.experts.92", "model.encoder.language_model.layers.13.experts.93", "model.encoder.language_model.layers.13.experts.94", "model.encoder.language_model.layers.13.experts.95", "model.encoder.language_model.layers.13.experts.96", "model.encoder.language_model.layers.13.experts.97", "model.encoder.language_model.layers.13.experts.98", "model.encoder.language_model.layers.13.experts.99", "model.encoder.language_model.layers.13.experts.100", "model.encoder.language_model.layers.13.experts.101", "model.encoder.language_model.layers.13.experts.102", "model.encoder.language_model.layers.13.experts.103", "model.encoder.language_model.layers.13.experts.104", "model.encoder.language_model.layers.13.experts.105", "model.encoder.language_model.layers.13.experts.106", "model.encoder.language_model.layers.13.experts.107", "model.encoder.language_model.layers.13.experts.108", "model.encoder.language_model.layers.13.experts.109", "model.encoder.language_model.layers.13.experts.110", "model.encoder.language_model.layers.13.experts.111", "model.encoder.language_model.layers.13.experts.112", "model.encoder.language_model.layers.13.experts.113", "model.encoder.language_model.layers.13.experts.114", "model.encoder.language_model.layers.13.experts.115", "model.encoder.language_model.layers.13.experts.116", "model.encoder.language_model.layers.13.experts.117", "model.encoder.language_model.layers.13.experts.118", "model.encoder.language_model.layers.13.experts.119", "model.encoder.language_model.layers.13.experts.120", "model.encoder.language_model.layers.13.experts.121", "model.encoder.language_model.layers.13.experts.122", "model.encoder.language_model.layers.13.experts.123", "model.encoder.language_model.layers.13.experts.124", "model.encoder.language_model.layers.13.experts.125", "model.encoder.language_model.layers.13.experts.126", "model.encoder.language_model.layers.13.experts.127", "model.encoder.language_model.layers.14.router", "model.encoder.language_model.layers.14.router.proj", "model.encoder.language_model.layers.14.experts.0", "model.encoder.language_model.layers.14.experts.1", "model.encoder.language_model.layers.14.experts.2", "model.encoder.language_model.layers.14.experts.3", "model.encoder.language_model.layers.14.experts.4", "model.encoder.language_model.layers.14.experts.5", "model.encoder.language_model.layers.14.experts.6", "model.encoder.language_model.layers.14.experts.7", "model.encoder.language_model.layers.14.experts.8", "model.encoder.language_model.layers.14.experts.9", "model.encoder.language_model.layers.14.experts.10", "model.encoder.language_model.layers.14.experts.11", "model.encoder.language_model.layers.14.experts.12", "model.encoder.language_model.layers.14.experts.13", "model.encoder.language_model.layers.14.experts.14", "model.encoder.language_model.layers.14.experts.15", "model.encoder.language_model.layers.14.experts.16", "model.encoder.language_model.layers.14.experts.17", "model.encoder.language_model.layers.14.experts.18", "model.encoder.language_model.layers.14.experts.19", "model.encoder.language_model.layers.14.experts.20", "model.encoder.language_model.layers.14.experts.21", "model.encoder.language_model.layers.14.experts.22", "model.encoder.language_model.layers.14.experts.23", "model.encoder.language_model.layers.14.experts.24", "model.encoder.language_model.layers.14.experts.25", "model.encoder.language_model.layers.14.experts.26", "model.encoder.language_model.layers.14.experts.27", "model.encoder.language_model.layers.14.experts.28", "model.encoder.language_model.layers.14.experts.29", "model.encoder.language_model.layers.14.experts.30", "model.encoder.language_model.layers.14.experts.31", "model.encoder.language_model.layers.14.experts.32", "model.encoder.language_model.layers.14.experts.33", "model.encoder.language_model.layers.14.experts.34", "model.encoder.language_model.layers.14.experts.35", "model.encoder.language_model.layers.14.experts.36", "model.encoder.language_model.layers.14.experts.37", "model.encoder.language_model.layers.14.experts.38", "model.encoder.language_model.layers.14.experts.39", "model.encoder.language_model.layers.14.experts.40", "model.encoder.language_model.layers.14.experts.41", "model.encoder.language_model.layers.14.experts.42", "model.encoder.language_model.layers.14.experts.43", "model.encoder.language_model.layers.14.experts.44", "model.encoder.language_model.layers.14.experts.45", "model.encoder.language_model.layers.14.experts.46", "model.encoder.language_model.layers.14.experts.47", "model.encoder.language_model.layers.14.experts.48", "model.encoder.language_model.layers.14.experts.49", "model.encoder.language_model.layers.14.experts.50", "model.encoder.language_model.layers.14.experts.51", "model.encoder.language_model.layers.14.experts.52", "model.encoder.language_model.layers.14.experts.53", "model.encoder.language_model.layers.14.experts.54", "model.encoder.language_model.layers.14.experts.55", "model.encoder.language_model.layers.14.experts.56", "model.encoder.language_model.layers.14.experts.57", "model.encoder.language_model.layers.14.experts.58", "model.encoder.language_model.layers.14.experts.59", "model.encoder.language_model.layers.14.experts.60", "model.encoder.language_model.layers.14.experts.61", "model.encoder.language_model.layers.14.experts.62", "model.encoder.language_model.layers.14.experts.63", "model.encoder.language_model.layers.14.experts.64", "model.encoder.language_model.layers.14.experts.65", "model.encoder.language_model.layers.14.experts.66", "model.encoder.language_model.layers.14.experts.67", "model.encoder.language_model.layers.14.experts.68", "model.encoder.language_model.layers.14.experts.69", "model.encoder.language_model.layers.14.experts.70", "model.encoder.language_model.layers.14.experts.71", "model.encoder.language_model.layers.14.experts.72", "model.encoder.language_model.layers.14.experts.73", "model.encoder.language_model.layers.14.experts.74", "model.encoder.language_model.layers.14.experts.75", "model.encoder.language_model.layers.14.experts.76", "model.encoder.language_model.layers.14.experts.77", "model.encoder.language_model.layers.14.experts.78", "model.encoder.language_model.layers.14.experts.79", "model.encoder.language_model.layers.14.experts.80", "model.encoder.language_model.layers.14.experts.81", "model.encoder.language_model.layers.14.experts.82", "model.encoder.language_model.layers.14.experts.83", "model.encoder.language_model.layers.14.experts.84", "model.encoder.language_model.layers.14.experts.85", "model.encoder.language_model.layers.14.experts.86", "model.encoder.language_model.layers.14.experts.87", "model.encoder.language_model.layers.14.experts.88", "model.encoder.language_model.layers.14.experts.89", "model.encoder.language_model.layers.14.experts.90", "model.encoder.language_model.layers.14.experts.91", "model.encoder.language_model.layers.14.experts.92", "model.encoder.language_model.layers.14.experts.93", "model.encoder.language_model.layers.14.experts.94", "model.encoder.language_model.layers.14.experts.95", "model.encoder.language_model.layers.14.experts.96", "model.encoder.language_model.layers.14.experts.97", "model.encoder.language_model.layers.14.experts.98", "model.encoder.language_model.layers.14.experts.99", "model.encoder.language_model.layers.14.experts.100", "model.encoder.language_model.layers.14.experts.101", "model.encoder.language_model.layers.14.experts.102", "model.encoder.language_model.layers.14.experts.103", "model.encoder.language_model.layers.14.experts.104", "model.encoder.language_model.layers.14.experts.105", "model.encoder.language_model.layers.14.experts.106", "model.encoder.language_model.layers.14.experts.107", "model.encoder.language_model.layers.14.experts.108", "model.encoder.language_model.layers.14.experts.109", "model.encoder.language_model.layers.14.experts.110", "model.encoder.language_model.layers.14.experts.111", "model.encoder.language_model.layers.14.experts.112", "model.encoder.language_model.layers.14.experts.113", "model.encoder.language_model.layers.14.experts.114", "model.encoder.language_model.layers.14.experts.115", "model.encoder.language_model.layers.14.experts.116", "model.encoder.language_model.layers.14.experts.117", "model.encoder.language_model.layers.14.experts.118", "model.encoder.language_model.layers.14.experts.119", "model.encoder.language_model.layers.14.experts.120", "model.encoder.language_model.layers.14.experts.121", "model.encoder.language_model.layers.14.experts.122", "model.encoder.language_model.layers.14.experts.123", "model.encoder.language_model.layers.14.experts.124", "model.encoder.language_model.layers.14.experts.125", "model.encoder.language_model.layers.14.experts.126", "model.encoder.language_model.layers.14.experts.127", "model.encoder.language_model.layers.15.router", "model.encoder.language_model.layers.15.router.proj", "model.encoder.language_model.layers.15.experts.0", "model.encoder.language_model.layers.15.experts.1", "model.encoder.language_model.layers.15.experts.2", "model.encoder.language_model.layers.15.experts.3", "model.encoder.language_model.layers.15.experts.4", "model.encoder.language_model.layers.15.experts.5", "model.encoder.language_model.layers.15.experts.6", "model.encoder.language_model.layers.15.experts.7", "model.encoder.language_model.layers.15.experts.8", "model.encoder.language_model.layers.15.experts.9", "model.encoder.language_model.layers.15.experts.10", "model.encoder.language_model.layers.15.experts.11", "model.encoder.language_model.layers.15.experts.12", "model.encoder.language_model.layers.15.experts.13", "model.encoder.language_model.layers.15.experts.14", "model.encoder.language_model.layers.15.experts.15", "model.encoder.language_model.layers.15.experts.16", "model.encoder.language_model.layers.15.experts.17", "model.encoder.language_model.layers.15.experts.18", "model.encoder.language_model.layers.15.experts.19", "model.encoder.language_model.layers.15.experts.20", "model.encoder.language_model.layers.15.experts.21", "model.encoder.language_model.layers.15.experts.22", "model.encoder.language_model.layers.15.experts.23", "model.encoder.language_model.layers.15.experts.24", "model.encoder.language_model.layers.15.experts.25", "model.encoder.language_model.layers.15.experts.26", "model.encoder.language_model.layers.15.experts.27", "model.encoder.language_model.layers.15.experts.28", "model.encoder.language_model.layers.15.experts.29", "model.encoder.language_model.layers.15.experts.30", "model.encoder.language_model.layers.15.experts.31", "model.encoder.language_model.layers.15.experts.32", "model.encoder.language_model.layers.15.experts.33", "model.encoder.language_model.layers.15.experts.34", "model.encoder.language_model.layers.15.experts.35", "model.encoder.language_model.layers.15.experts.36", "model.encoder.language_model.layers.15.experts.37", "model.encoder.language_model.layers.15.experts.38", "model.encoder.language_model.layers.15.experts.39", "model.encoder.language_model.layers.15.experts.40", "model.encoder.language_model.layers.15.experts.41", "model.encoder.language_model.layers.15.experts.42", "model.encoder.language_model.layers.15.experts.43", "model.encoder.language_model.layers.15.experts.44", "model.encoder.language_model.layers.15.experts.45", "model.encoder.language_model.layers.15.experts.46", "model.encoder.language_model.layers.15.experts.47", "model.encoder.language_model.layers.15.experts.48", "model.encoder.language_model.layers.15.experts.49", "model.encoder.language_model.layers.15.experts.50", "model.encoder.language_model.layers.15.experts.51", "model.encoder.language_model.layers.15.experts.52", "model.encoder.language_model.layers.15.experts.53", "model.encoder.language_model.layers.15.experts.54", "model.encoder.language_model.layers.15.experts.55", "model.encoder.language_model.layers.15.experts.56", "model.encoder.language_model.layers.15.experts.57", "model.encoder.language_model.layers.15.experts.58", "model.encoder.language_model.layers.15.experts.59", "model.encoder.language_model.layers.15.experts.60", "model.encoder.language_model.layers.15.experts.61", "model.encoder.language_model.layers.15.experts.62", "model.encoder.language_model.layers.15.experts.63", "model.encoder.language_model.layers.15.experts.64", "model.encoder.language_model.layers.15.experts.65", "model.encoder.language_model.layers.15.experts.66", "model.encoder.language_model.layers.15.experts.67", "model.encoder.language_model.layers.15.experts.68", "model.encoder.language_model.layers.15.experts.69", "model.encoder.language_model.layers.15.experts.70", "model.encoder.language_model.layers.15.experts.71", "model.encoder.language_model.layers.15.experts.72", "model.encoder.language_model.layers.15.experts.73", "model.encoder.language_model.layers.15.experts.74", "model.encoder.language_model.layers.15.experts.75", "model.encoder.language_model.layers.15.experts.76", "model.encoder.language_model.layers.15.experts.77", "model.encoder.language_model.layers.15.experts.78", "model.encoder.language_model.layers.15.experts.79", "model.encoder.language_model.layers.15.experts.80", "model.encoder.language_model.layers.15.experts.81", "model.encoder.language_model.layers.15.experts.82", "model.encoder.language_model.layers.15.experts.83", "model.encoder.language_model.layers.15.experts.84", "model.encoder.language_model.layers.15.experts.85", "model.encoder.language_model.layers.15.experts.86", "model.encoder.language_model.layers.15.experts.87", "model.encoder.language_model.layers.15.experts.88", "model.encoder.language_model.layers.15.experts.89", "model.encoder.language_model.layers.15.experts.90", "model.encoder.language_model.layers.15.experts.91", "model.encoder.language_model.layers.15.experts.92", "model.encoder.language_model.layers.15.experts.93", "model.encoder.language_model.layers.15.experts.94", "model.encoder.language_model.layers.15.experts.95", "model.encoder.language_model.layers.15.experts.96", "model.encoder.language_model.layers.15.experts.97", "model.encoder.language_model.layers.15.experts.98", "model.encoder.language_model.layers.15.experts.99", "model.encoder.language_model.layers.15.experts.100", "model.encoder.language_model.layers.15.experts.101", "model.encoder.language_model.layers.15.experts.102", "model.encoder.language_model.layers.15.experts.103", "model.encoder.language_model.layers.15.experts.104", "model.encoder.language_model.layers.15.experts.105", "model.encoder.language_model.layers.15.experts.106", "model.encoder.language_model.layers.15.experts.107", "model.encoder.language_model.layers.15.experts.108", "model.encoder.language_model.layers.15.experts.109", "model.encoder.language_model.layers.15.experts.110", "model.encoder.language_model.layers.15.experts.111", "model.encoder.language_model.layers.15.experts.112", "model.encoder.language_model.layers.15.experts.113", "model.encoder.language_model.layers.15.experts.114", "model.encoder.language_model.layers.15.experts.115", "model.encoder.language_model.layers.15.experts.116", "model.encoder.language_model.layers.15.experts.117", "model.encoder.language_model.layers.15.experts.118", "model.encoder.language_model.layers.15.experts.119", "model.encoder.language_model.layers.15.experts.120", "model.encoder.language_model.layers.15.experts.121", "model.encoder.language_model.layers.15.experts.122", "model.encoder.language_model.layers.15.experts.123", "model.encoder.language_model.layers.15.experts.124", "model.encoder.language_model.layers.15.experts.125", "model.encoder.language_model.layers.15.experts.126", "model.encoder.language_model.layers.15.experts.127", "model.encoder.language_model.layers.16.router", "model.encoder.language_model.layers.16.router.proj", "model.encoder.language_model.layers.16.experts.0", "model.encoder.language_model.layers.16.experts.1", "model.encoder.language_model.layers.16.experts.2", "model.encoder.language_model.layers.16.experts.3", "model.encoder.language_model.layers.16.experts.4", "model.encoder.language_model.layers.16.experts.5", "model.encoder.language_model.layers.16.experts.6", "model.encoder.language_model.layers.16.experts.7", "model.encoder.language_model.layers.16.experts.8", "model.encoder.language_model.layers.16.experts.9", "model.encoder.language_model.layers.16.experts.10", "model.encoder.language_model.layers.16.experts.11", "model.encoder.language_model.layers.16.experts.12", "model.encoder.language_model.layers.16.experts.13", "model.encoder.language_model.layers.16.experts.14", "model.encoder.language_model.layers.16.experts.15", "model.encoder.language_model.layers.16.experts.16", "model.encoder.language_model.layers.16.experts.17", "model.encoder.language_model.layers.16.experts.18", "model.encoder.language_model.layers.16.experts.19", "model.encoder.language_model.layers.16.experts.20", "model.encoder.language_model.layers.16.experts.21", "model.encoder.language_model.layers.16.experts.22", "model.encoder.language_model.layers.16.experts.23", "model.encoder.language_model.layers.16.experts.24", "model.encoder.language_model.layers.16.experts.25", "model.encoder.language_model.layers.16.experts.26", "model.encoder.language_model.layers.16.experts.27", "model.encoder.language_model.layers.16.experts.28", "model.encoder.language_model.layers.16.experts.29", "model.encoder.language_model.layers.16.experts.30", "model.encoder.language_model.layers.16.experts.31", "model.encoder.language_model.layers.16.experts.32", "model.encoder.language_model.layers.16.experts.33", "model.encoder.language_model.layers.16.experts.34", "model.encoder.language_model.layers.16.experts.35", "model.encoder.language_model.layers.16.experts.36", "model.encoder.language_model.layers.16.experts.37", "model.encoder.language_model.layers.16.experts.38", "model.encoder.language_model.layers.16.experts.39", "model.encoder.language_model.layers.16.experts.40", "model.encoder.language_model.layers.16.experts.41", "model.encoder.language_model.layers.16.experts.42", "model.encoder.language_model.layers.16.experts.43", "model.encoder.language_model.layers.16.experts.44", "model.encoder.language_model.layers.16.experts.45", "model.encoder.language_model.layers.16.experts.46", "model.encoder.language_model.layers.16.experts.47", "model.encoder.language_model.layers.16.experts.48", "model.encoder.language_model.layers.16.experts.49", "model.encoder.language_model.layers.16.experts.50", "model.encoder.language_model.layers.16.experts.51", "model.encoder.language_model.layers.16.experts.52", "model.encoder.language_model.layers.16.experts.53", "model.encoder.language_model.layers.16.experts.54", "model.encoder.language_model.layers.16.experts.55", "model.encoder.language_model.layers.16.experts.56", "model.encoder.language_model.layers.16.experts.57", "model.encoder.language_model.layers.16.experts.58", "model.encoder.language_model.layers.16.experts.59", "model.encoder.language_model.layers.16.experts.60", "model.encoder.language_model.layers.16.experts.61", "model.encoder.language_model.layers.16.experts.62", "model.encoder.language_model.layers.16.experts.63", "model.encoder.language_model.layers.16.experts.64", "model.encoder.language_model.layers.16.experts.65", "model.encoder.language_model.layers.16.experts.66", "model.encoder.language_model.layers.16.experts.67", "model.encoder.language_model.layers.16.experts.68", "model.encoder.language_model.layers.16.experts.69", "model.encoder.language_model.layers.16.experts.70", "model.encoder.language_model.layers.16.experts.71", "model.encoder.language_model.layers.16.experts.72", "model.encoder.language_model.layers.16.experts.73", "model.encoder.language_model.layers.16.experts.74", "model.encoder.language_model.layers.16.experts.75", "model.encoder.language_model.layers.16.experts.76", "model.encoder.language_model.layers.16.experts.77", "model.encoder.language_model.layers.16.experts.78", "model.encoder.language_model.layers.16.experts.79", "model.encoder.language_model.layers.16.experts.80", "model.encoder.language_model.layers.16.experts.81", "model.encoder.language_model.layers.16.experts.82", "model.encoder.language_model.layers.16.experts.83", "model.encoder.language_model.layers.16.experts.84", "model.encoder.language_model.layers.16.experts.85", "model.encoder.language_model.layers.16.experts.86", "model.encoder.language_model.layers.16.experts.87", "model.encoder.language_model.layers.16.experts.88", "model.encoder.language_model.layers.16.experts.89", "model.encoder.language_model.layers.16.experts.90", "model.encoder.language_model.layers.16.experts.91", "model.encoder.language_model.layers.16.experts.92", "model.encoder.language_model.layers.16.experts.93", "model.encoder.language_model.layers.16.experts.94", "model.encoder.language_model.layers.16.experts.95", "model.encoder.language_model.layers.16.experts.96", "model.encoder.language_model.layers.16.experts.97", "model.encoder.language_model.layers.16.experts.98", "model.encoder.language_model.layers.16.experts.99", "model.encoder.language_model.layers.16.experts.100", "model.encoder.language_model.layers.16.experts.101", "model.encoder.language_model.layers.16.experts.102", "model.encoder.language_model.layers.16.experts.103", "model.encoder.language_model.layers.16.experts.104", "model.encoder.language_model.layers.16.experts.105", "model.encoder.language_model.layers.16.experts.106", "model.encoder.language_model.layers.16.experts.107", "model.encoder.language_model.layers.16.experts.108", "model.encoder.language_model.layers.16.experts.109", "model.encoder.language_model.layers.16.experts.110", "model.encoder.language_model.layers.16.experts.111", "model.encoder.language_model.layers.16.experts.112", "model.encoder.language_model.layers.16.experts.113", "model.encoder.language_model.layers.16.experts.114", "model.encoder.language_model.layers.16.experts.115", "model.encoder.language_model.layers.16.experts.116", "model.encoder.language_model.layers.16.experts.117", "model.encoder.language_model.layers.16.experts.118", "model.encoder.language_model.layers.16.experts.119", "model.encoder.language_model.layers.16.experts.120", "model.encoder.language_model.layers.16.experts.121", "model.encoder.language_model.layers.16.experts.122", "model.encoder.language_model.layers.16.experts.123", "model.encoder.language_model.layers.16.experts.124", "model.encoder.language_model.layers.16.experts.125", "model.encoder.language_model.layers.16.experts.126", "model.encoder.language_model.layers.16.experts.127", "model.encoder.language_model.layers.17.router", "model.encoder.language_model.layers.17.router.proj", "model.encoder.language_model.layers.17.experts.0", "model.encoder.language_model.layers.17.experts.1", "model.encoder.language_model.layers.17.experts.2", "model.encoder.language_model.layers.17.experts.3", "model.encoder.language_model.layers.17.experts.4", "model.encoder.language_model.layers.17.experts.5", "model.encoder.language_model.layers.17.experts.6", "model.encoder.language_model.layers.17.experts.7", "model.encoder.language_model.layers.17.experts.8", "model.encoder.language_model.layers.17.experts.9", "model.encoder.language_model.layers.17.experts.10", "model.encoder.language_model.layers.17.experts.11", "model.encoder.language_model.layers.17.experts.12", "model.encoder.language_model.layers.17.experts.13", "model.encoder.language_model.layers.17.experts.14", "model.encoder.language_model.layers.17.experts.15", "model.encoder.language_model.layers.17.experts.16", "model.encoder.language_model.layers.17.experts.17", "model.encoder.language_model.layers.17.experts.18", "model.encoder.language_model.layers.17.experts.19", "model.encoder.language_model.layers.17.experts.20", "model.encoder.language_model.layers.17.experts.21", "model.encoder.language_model.layers.17.experts.22", "model.encoder.language_model.layers.17.experts.23", "model.encoder.language_model.layers.17.experts.24", "model.encoder.language_model.layers.17.experts.25", "model.encoder.language_model.layers.17.experts.26", "model.encoder.language_model.layers.17.experts.27", "model.encoder.language_model.layers.17.experts.28", "model.encoder.language_model.layers.17.experts.29", "model.encoder.language_model.layers.17.experts.30", "model.encoder.language_model.layers.17.experts.31", "model.encoder.language_model.layers.17.experts.32", "model.encoder.language_model.layers.17.experts.33", "model.encoder.language_model.layers.17.experts.34", "model.encoder.language_model.layers.17.experts.35", "model.encoder.language_model.layers.17.experts.36", "model.encoder.language_model.layers.17.experts.37", "model.encoder.language_model.layers.17.experts.38", "model.encoder.language_model.layers.17.experts.39", "model.encoder.language_model.layers.17.experts.40", "model.encoder.language_model.layers.17.experts.41", "model.encoder.language_model.layers.17.experts.42", "model.encoder.language_model.layers.17.experts.43", "model.encoder.language_model.layers.17.experts.44", "model.encoder.language_model.layers.17.experts.45", "model.encoder.language_model.layers.17.experts.46", "model.encoder.language_model.layers.17.experts.47", "model.encoder.language_model.layers.17.experts.48", "model.encoder.language_model.layers.17.experts.49", "model.encoder.language_model.layers.17.experts.50", "model.encoder.language_model.layers.17.experts.51", "model.encoder.language_model.layers.17.experts.52", "model.encoder.language_model.layers.17.experts.53", "model.encoder.language_model.layers.17.experts.54", "model.encoder.language_model.layers.17.experts.55", "model.encoder.language_model.layers.17.experts.56", "model.encoder.language_model.layers.17.experts.57", "model.encoder.language_model.layers.17.experts.58", "model.encoder.language_model.layers.17.experts.59", "model.encoder.language_model.layers.17.experts.60", "model.encoder.language_model.layers.17.experts.61", "model.encoder.language_model.layers.17.experts.62", "model.encoder.language_model.layers.17.experts.63", "model.encoder.language_model.layers.17.experts.64", "model.encoder.language_model.layers.17.experts.65", "model.encoder.language_model.layers.17.experts.66", "model.encoder.language_model.layers.17.experts.67", "model.encoder.language_model.layers.17.experts.68", "model.encoder.language_model.layers.17.experts.69", "model.encoder.language_model.layers.17.experts.70", "model.encoder.language_model.layers.17.experts.71", "model.encoder.language_model.layers.17.experts.72", "model.encoder.language_model.layers.17.experts.73", "model.encoder.language_model.layers.17.experts.74", "model.encoder.language_model.layers.17.experts.75", "model.encoder.language_model.layers.17.experts.76", "model.encoder.language_model.layers.17.experts.77", "model.encoder.language_model.layers.17.experts.78", "model.encoder.language_model.layers.17.experts.79", "model.encoder.language_model.layers.17.experts.80", "model.encoder.language_model.layers.17.experts.81", "model.encoder.language_model.layers.17.experts.82", "model.encoder.language_model.layers.17.experts.83", "model.encoder.language_model.layers.17.experts.84", "model.encoder.language_model.layers.17.experts.85", "model.encoder.language_model.layers.17.experts.86", "model.encoder.language_model.layers.17.experts.87", "model.encoder.language_model.layers.17.experts.88", "model.encoder.language_model.layers.17.experts.89", "model.encoder.language_model.layers.17.experts.90", "model.encoder.language_model.layers.17.experts.91", "model.encoder.language_model.layers.17.experts.92", "model.encoder.language_model.layers.17.experts.93", "model.encoder.language_model.layers.17.experts.94", "model.encoder.language_model.layers.17.experts.95", "model.encoder.language_model.layers.17.experts.96", "model.encoder.language_model.layers.17.experts.97", "model.encoder.language_model.layers.17.experts.98", "model.encoder.language_model.layers.17.experts.99", "model.encoder.language_model.layers.17.experts.100", "model.encoder.language_model.layers.17.experts.101", "model.encoder.language_model.layers.17.experts.102", "model.encoder.language_model.layers.17.experts.103", "model.encoder.language_model.layers.17.experts.104", "model.encoder.language_model.layers.17.experts.105", "model.encoder.language_model.layers.17.experts.106", "model.encoder.language_model.layers.17.experts.107", "model.encoder.language_model.layers.17.experts.108", "model.encoder.language_model.layers.17.experts.109", "model.encoder.language_model.layers.17.experts.110", "model.encoder.language_model.layers.17.experts.111", "model.encoder.language_model.layers.17.experts.112", "model.encoder.language_model.layers.17.experts.113", "model.encoder.language_model.layers.17.experts.114", "model.encoder.language_model.layers.17.experts.115", "model.encoder.language_model.layers.17.experts.116", "model.encoder.language_model.layers.17.experts.117", "model.encoder.language_model.layers.17.experts.118", "model.encoder.language_model.layers.17.experts.119", "model.encoder.language_model.layers.17.experts.120", "model.encoder.language_model.layers.17.experts.121", "model.encoder.language_model.layers.17.experts.122", "model.encoder.language_model.layers.17.experts.123", "model.encoder.language_model.layers.17.experts.124", "model.encoder.language_model.layers.17.experts.125", "model.encoder.language_model.layers.17.experts.126", "model.encoder.language_model.layers.17.experts.127", "model.encoder.language_model.layers.18.router", "model.encoder.language_model.layers.18.router.proj", "model.encoder.language_model.layers.18.experts.0", "model.encoder.language_model.layers.18.experts.1", "model.encoder.language_model.layers.18.experts.2", "model.encoder.language_model.layers.18.experts.3", "model.encoder.language_model.layers.18.experts.4", "model.encoder.language_model.layers.18.experts.5", "model.encoder.language_model.layers.18.experts.6", "model.encoder.language_model.layers.18.experts.7", "model.encoder.language_model.layers.18.experts.8", "model.encoder.language_model.layers.18.experts.9", "model.encoder.language_model.layers.18.experts.10", "model.encoder.language_model.layers.18.experts.11", "model.encoder.language_model.layers.18.experts.12", "model.encoder.language_model.layers.18.experts.13", "model.encoder.language_model.layers.18.experts.14", "model.encoder.language_model.layers.18.experts.15", "model.encoder.language_model.layers.18.experts.16", "model.encoder.language_model.layers.18.experts.17", "model.encoder.language_model.layers.18.experts.18", "model.encoder.language_model.layers.18.experts.19", "model.encoder.language_model.layers.18.experts.20", "model.encoder.language_model.layers.18.experts.21", "model.encoder.language_model.layers.18.experts.22", "model.encoder.language_model.layers.18.experts.23", "model.encoder.language_model.layers.18.experts.24", "model.encoder.language_model.layers.18.experts.25", "model.encoder.language_model.layers.18.experts.26", "model.encoder.language_model.layers.18.experts.27", "model.encoder.language_model.layers.18.experts.28", "model.encoder.language_model.layers.18.experts.29", "model.encoder.language_model.layers.18.experts.30", "model.encoder.language_model.layers.18.experts.31", "model.encoder.language_model.layers.18.experts.32", "model.encoder.language_model.layers.18.experts.33", "model.encoder.language_model.layers.18.experts.34", "model.encoder.language_model.layers.18.experts.35", "model.encoder.language_model.layers.18.experts.36", "model.encoder.language_model.layers.18.experts.37", "model.encoder.language_model.layers.18.experts.38", "model.encoder.language_model.layers.18.experts.39", "model.encoder.language_model.layers.18.experts.40", "model.encoder.language_model.layers.18.experts.41", "model.encoder.language_model.layers.18.experts.42", "model.encoder.language_model.layers.18.experts.43", "model.encoder.language_model.layers.18.experts.44", "model.encoder.language_model.layers.18.experts.45", "model.encoder.language_model.layers.18.experts.46", "model.encoder.language_model.layers.18.experts.47", "model.encoder.language_model.layers.18.experts.48", "model.encoder.language_model.layers.18.experts.49", "model.encoder.language_model.layers.18.experts.50", "model.encoder.language_model.layers.18.experts.51", "model.encoder.language_model.layers.18.experts.52", "model.encoder.language_model.layers.18.experts.53", "model.encoder.language_model.layers.18.experts.54", "model.encoder.language_model.layers.18.experts.55", "model.encoder.language_model.layers.18.experts.56", "model.encoder.language_model.layers.18.experts.57", "model.encoder.language_model.layers.18.experts.58", "model.encoder.language_model.layers.18.experts.59", "model.encoder.language_model.layers.18.experts.60", "model.encoder.language_model.layers.18.experts.61", "model.encoder.language_model.layers.18.experts.62", "model.encoder.language_model.layers.18.experts.63", "model.encoder.language_model.layers.18.experts.64", "model.encoder.language_model.layers.18.experts.65", "model.encoder.language_model.layers.18.experts.66", "model.encoder.language_model.layers.18.experts.67", "model.encoder.language_model.layers.18.experts.68", "model.encoder.language_model.layers.18.experts.69", "model.encoder.language_model.layers.18.experts.70", "model.encoder.language_model.layers.18.experts.71", "model.encoder.language_model.layers.18.experts.72", "model.encoder.language_model.layers.18.experts.73", "model.encoder.language_model.layers.18.experts.74", "model.encoder.language_model.layers.18.experts.75", "model.encoder.language_model.layers.18.experts.76", "model.encoder.language_model.layers.18.experts.77", "model.encoder.language_model.layers.18.experts.78", "model.encoder.language_model.layers.18.experts.79", "model.encoder.language_model.layers.18.experts.80", "model.encoder.language_model.layers.18.experts.81", "model.encoder.language_model.layers.18.experts.82", "model.encoder.language_model.layers.18.experts.83", "model.encoder.language_model.layers.18.experts.84", "model.encoder.language_model.layers.18.experts.85", "model.encoder.language_model.layers.18.experts.86", "model.encoder.language_model.layers.18.experts.87", "model.encoder.language_model.layers.18.experts.88", "model.encoder.language_model.layers.18.experts.89", "model.encoder.language_model.layers.18.experts.90", "model.encoder.language_model.layers.18.experts.91", "model.encoder.language_model.layers.18.experts.92", "model.encoder.language_model.layers.18.experts.93", "model.encoder.language_model.layers.18.experts.94", "model.encoder.language_model.layers.18.experts.95", "model.encoder.language_model.layers.18.experts.96", "model.encoder.language_model.layers.18.experts.97", "model.encoder.language_model.layers.18.experts.98", "model.encoder.language_model.layers.18.experts.99", "model.encoder.language_model.layers.18.experts.100", "model.encoder.language_model.layers.18.experts.101", "model.encoder.language_model.layers.18.experts.102", "model.encoder.language_model.layers.18.experts.103", "model.encoder.language_model.layers.18.experts.104", "model.encoder.language_model.layers.18.experts.105", "model.encoder.language_model.layers.18.experts.106", "model.encoder.language_model.layers.18.experts.107", "model.encoder.language_model.layers.18.experts.108", "model.encoder.language_model.layers.18.experts.109", "model.encoder.language_model.layers.18.experts.110", "model.encoder.language_model.layers.18.experts.111", "model.encoder.language_model.layers.18.experts.112", "model.encoder.language_model.layers.18.experts.113", "model.encoder.language_model.layers.18.experts.114", "model.encoder.language_model.layers.18.experts.115", "model.encoder.language_model.layers.18.experts.116", "model.encoder.language_model.layers.18.experts.117", "model.encoder.language_model.layers.18.experts.118", "model.encoder.language_model.layers.18.experts.119", "model.encoder.language_model.layers.18.experts.120", "model.encoder.language_model.layers.18.experts.121", "model.encoder.language_model.layers.18.experts.122", "model.encoder.language_model.layers.18.experts.123", "model.encoder.language_model.layers.18.experts.124", "model.encoder.language_model.layers.18.experts.125", "model.encoder.language_model.layers.18.experts.126", "model.encoder.language_model.layers.18.experts.127", "model.encoder.language_model.layers.19.router", "model.encoder.language_model.layers.19.router.proj", "model.encoder.language_model.layers.19.experts.0", "model.encoder.language_model.layers.19.experts.1", "model.encoder.language_model.layers.19.experts.2", "model.encoder.language_model.layers.19.experts.3", "model.encoder.language_model.layers.19.experts.4", "model.encoder.language_model.layers.19.experts.5", "model.encoder.language_model.layers.19.experts.6", "model.encoder.language_model.layers.19.experts.7", "model.encoder.language_model.layers.19.experts.8", "model.encoder.language_model.layers.19.experts.9", "model.encoder.language_model.layers.19.experts.10", "model.encoder.language_model.layers.19.experts.11", "model.encoder.language_model.layers.19.experts.12", "model.encoder.language_model.layers.19.experts.13", "model.encoder.language_model.layers.19.experts.14", "model.encoder.language_model.layers.19.experts.15", "model.encoder.language_model.layers.19.experts.16", "model.encoder.language_model.layers.19.experts.17", "model.encoder.language_model.layers.19.experts.18", "model.encoder.language_model.layers.19.experts.19", "model.encoder.language_model.layers.19.experts.20", "model.encoder.language_model.layers.19.experts.21", "model.encoder.language_model.layers.19.experts.22", "model.encoder.language_model.layers.19.experts.23", "model.encoder.language_model.layers.19.experts.24", "model.encoder.language_model.layers.19.experts.25", "model.encoder.language_model.layers.19.experts.26", "model.encoder.language_model.layers.19.experts.27", "model.encoder.language_model.layers.19.experts.28", "model.encoder.language_model.layers.19.experts.29", "model.encoder.language_model.layers.19.experts.30", "model.encoder.language_model.layers.19.experts.31", "model.encoder.language_model.layers.19.experts.32", "model.encoder.language_model.layers.19.experts.33", "model.encoder.language_model.layers.19.experts.34", "model.encoder.language_model.layers.19.experts.35", "model.encoder.language_model.layers.19.experts.36", "model.encoder.language_model.layers.19.experts.37", "model.encoder.language_model.layers.19.experts.38", "model.encoder.language_model.layers.19.experts.39", "model.encoder.language_model.layers.19.experts.40", "model.encoder.language_model.layers.19.experts.41", "model.encoder.language_model.layers.19.experts.42", "model.encoder.language_model.layers.19.experts.43", "model.encoder.language_model.layers.19.experts.44", "model.encoder.language_model.layers.19.experts.45", "model.encoder.language_model.layers.19.experts.46", "model.encoder.language_model.layers.19.experts.47", "model.encoder.language_model.layers.19.experts.48", "model.encoder.language_model.layers.19.experts.49", "model.encoder.language_model.layers.19.experts.50", "model.encoder.language_model.layers.19.experts.51", "model.encoder.language_model.layers.19.experts.52", "model.encoder.language_model.layers.19.experts.53", "model.encoder.language_model.layers.19.experts.54", "model.encoder.language_model.layers.19.experts.55", "model.encoder.language_model.layers.19.experts.56", "model.encoder.language_model.layers.19.experts.57", "model.encoder.language_model.layers.19.experts.58", "model.encoder.language_model.layers.19.experts.59", "model.encoder.language_model.layers.19.experts.60", "model.encoder.language_model.layers.19.experts.61", "model.encoder.language_model.layers.19.experts.62", "model.encoder.language_model.layers.19.experts.63", "model.encoder.language_model.layers.19.experts.64", "model.encoder.language_model.layers.19.experts.65", "model.encoder.language_model.layers.19.experts.66", "model.encoder.language_model.layers.19.experts.67", "model.encoder.language_model.layers.19.experts.68", "model.encoder.language_model.layers.19.experts.69", "model.encoder.language_model.layers.19.experts.70", "model.encoder.language_model.layers.19.experts.71", "model.encoder.language_model.layers.19.experts.72", "model.encoder.language_model.layers.19.experts.73", "model.encoder.language_model.layers.19.experts.74", "model.encoder.language_model.layers.19.experts.75", "model.encoder.language_model.layers.19.experts.76", "model.encoder.language_model.layers.19.experts.77", "model.encoder.language_model.layers.19.experts.78", "model.encoder.language_model.layers.19.experts.79", "model.encoder.language_model.layers.19.experts.80", "model.encoder.language_model.layers.19.experts.81", "model.encoder.language_model.layers.19.experts.82", "model.encoder.language_model.layers.19.experts.83", "model.encoder.language_model.layers.19.experts.84", "model.encoder.language_model.layers.19.experts.85", "model.encoder.language_model.layers.19.experts.86", "model.encoder.language_model.layers.19.experts.87", "model.encoder.language_model.layers.19.experts.88", "model.encoder.language_model.layers.19.experts.89", "model.encoder.language_model.layers.19.experts.90", "model.encoder.language_model.layers.19.experts.91", "model.encoder.language_model.layers.19.experts.92", "model.encoder.language_model.layers.19.experts.93", "model.encoder.language_model.layers.19.experts.94", "model.encoder.language_model.layers.19.experts.95", "model.encoder.language_model.layers.19.experts.96", "model.encoder.language_model.layers.19.experts.97", "model.encoder.language_model.layers.19.experts.98", "model.encoder.language_model.layers.19.experts.99", "model.encoder.language_model.layers.19.experts.100", "model.encoder.language_model.layers.19.experts.101", "model.encoder.language_model.layers.19.experts.102", "model.encoder.language_model.layers.19.experts.103", "model.encoder.language_model.layers.19.experts.104", "model.encoder.language_model.layers.19.experts.105", "model.encoder.language_model.layers.19.experts.106", "model.encoder.language_model.layers.19.experts.107", "model.encoder.language_model.layers.19.experts.108", "model.encoder.language_model.layers.19.experts.109", "model.encoder.language_model.layers.19.experts.110", "model.encoder.language_model.layers.19.experts.111", "model.encoder.language_model.layers.19.experts.112", "model.encoder.language_model.layers.19.experts.113", "model.encoder.language_model.layers.19.experts.114", "model.encoder.language_model.layers.19.experts.115", "model.encoder.language_model.layers.19.experts.116", "model.encoder.language_model.layers.19.experts.117", "model.encoder.language_model.layers.19.experts.118", "model.encoder.language_model.layers.19.experts.119", "model.encoder.language_model.layers.19.experts.120", "model.encoder.language_model.layers.19.experts.121", "model.encoder.language_model.layers.19.experts.122", "model.encoder.language_model.layers.19.experts.123", "model.encoder.language_model.layers.19.experts.124", "model.encoder.language_model.layers.19.experts.125", "model.encoder.language_model.layers.19.experts.126", "model.encoder.language_model.layers.19.experts.127", "model.encoder.language_model.layers.20.router", "model.encoder.language_model.layers.20.router.proj", "model.encoder.language_model.layers.20.experts.0", "model.encoder.language_model.layers.20.experts.1", "model.encoder.language_model.layers.20.experts.2", "model.encoder.language_model.layers.20.experts.3", "model.encoder.language_model.layers.20.experts.4", "model.encoder.language_model.layers.20.experts.5", "model.encoder.language_model.layers.20.experts.6", "model.encoder.language_model.layers.20.experts.7", "model.encoder.language_model.layers.20.experts.8", "model.encoder.language_model.layers.20.experts.9", "model.encoder.language_model.layers.20.experts.10", "model.encoder.language_model.layers.20.experts.11", "model.encoder.language_model.layers.20.experts.12", "model.encoder.language_model.layers.20.experts.13", "model.encoder.language_model.layers.20.experts.14", "model.encoder.language_model.layers.20.experts.15", "model.encoder.language_model.layers.20.experts.16", "model.encoder.language_model.layers.20.experts.17", "model.encoder.language_model.layers.20.experts.18", "model.encoder.language_model.layers.20.experts.19", "model.encoder.language_model.layers.20.experts.20", "model.encoder.language_model.layers.20.experts.21", "model.encoder.language_model.layers.20.experts.22", "model.encoder.language_model.layers.20.experts.23", "model.encoder.language_model.layers.20.experts.24", "model.encoder.language_model.layers.20.experts.25", "model.encoder.language_model.layers.20.experts.26", "model.encoder.language_model.layers.20.experts.27", "model.encoder.language_model.layers.20.experts.28", "model.encoder.language_model.layers.20.experts.29", "model.encoder.language_model.layers.20.experts.30", "model.encoder.language_model.layers.20.experts.31", "model.encoder.language_model.layers.20.experts.32", "model.encoder.language_model.layers.20.experts.33", "model.encoder.language_model.layers.20.experts.34", "model.encoder.language_model.layers.20.experts.35", "model.encoder.language_model.layers.20.experts.36", "model.encoder.language_model.layers.20.experts.37", "model.encoder.language_model.layers.20.experts.38", "model.encoder.language_model.layers.20.experts.39", "model.encoder.language_model.layers.20.experts.40", "model.encoder.language_model.layers.20.experts.41", "model.encoder.language_model.layers.20.experts.42", "model.encoder.language_model.layers.20.experts.43", "model.encoder.language_model.layers.20.experts.44", "model.encoder.language_model.layers.20.experts.45", "model.encoder.language_model.layers.20.experts.46", "model.encoder.language_model.layers.20.experts.47", "model.encoder.language_model.layers.20.experts.48", "model.encoder.language_model.layers.20.experts.49", "model.encoder.language_model.layers.20.experts.50", "model.encoder.language_model.layers.20.experts.51", "model.encoder.language_model.layers.20.experts.52", "model.encoder.language_model.layers.20.experts.53", "model.encoder.language_model.layers.20.experts.54", "model.encoder.language_model.layers.20.experts.55", "model.encoder.language_model.layers.20.experts.56", "model.encoder.language_model.layers.20.experts.57", "model.encoder.language_model.layers.20.experts.58", "model.encoder.language_model.layers.20.experts.59", "model.encoder.language_model.layers.20.experts.60", "model.encoder.language_model.layers.20.experts.61", "model.encoder.language_model.layers.20.experts.62", "model.encoder.language_model.layers.20.experts.63", "model.encoder.language_model.layers.20.experts.64", "model.encoder.language_model.layers.20.experts.65", "model.encoder.language_model.layers.20.experts.66", "model.encoder.language_model.layers.20.experts.67", "model.encoder.language_model.layers.20.experts.68", "model.encoder.language_model.layers.20.experts.69", "model.encoder.language_model.layers.20.experts.70", "model.encoder.language_model.layers.20.experts.71", "model.encoder.language_model.layers.20.experts.72", "model.encoder.language_model.layers.20.experts.73", "model.encoder.language_model.layers.20.experts.74", "model.encoder.language_model.layers.20.experts.75", "model.encoder.language_model.layers.20.experts.76", "model.encoder.language_model.layers.20.experts.77", "model.encoder.language_model.layers.20.experts.78", "model.encoder.language_model.layers.20.experts.79", "model.encoder.language_model.layers.20.experts.80", "model.encoder.language_model.layers.20.experts.81", "model.encoder.language_model.layers.20.experts.82", "model.encoder.language_model.layers.20.experts.83", "model.encoder.language_model.layers.20.experts.84", "model.encoder.language_model.layers.20.experts.85", "model.encoder.language_model.layers.20.experts.86", "model.encoder.language_model.layers.20.experts.87", "model.encoder.language_model.layers.20.experts.88", "model.encoder.language_model.layers.20.experts.89", "model.encoder.language_model.layers.20.experts.90", "model.encoder.language_model.layers.20.experts.91", "model.encoder.language_model.layers.20.experts.92", "model.encoder.language_model.layers.20.experts.93", "model.encoder.language_model.layers.20.experts.94", "model.encoder.language_model.layers.20.experts.95", "model.encoder.language_model.layers.20.experts.96", "model.encoder.language_model.layers.20.experts.97", "model.encoder.language_model.layers.20.experts.98", "model.encoder.language_model.layers.20.experts.99", "model.encoder.language_model.layers.20.experts.100", "model.encoder.language_model.layers.20.experts.101", "model.encoder.language_model.layers.20.experts.102", "model.encoder.language_model.layers.20.experts.103", "model.encoder.language_model.layers.20.experts.104", "model.encoder.language_model.layers.20.experts.105", "model.encoder.language_model.layers.20.experts.106", "model.encoder.language_model.layers.20.experts.107", "model.encoder.language_model.layers.20.experts.108", "model.encoder.language_model.layers.20.experts.109", "model.encoder.language_model.layers.20.experts.110", "model.encoder.language_model.layers.20.experts.111", "model.encoder.language_model.layers.20.experts.112", "model.encoder.language_model.layers.20.experts.113", "model.encoder.language_model.layers.20.experts.114", "model.encoder.language_model.layers.20.experts.115", "model.encoder.language_model.layers.20.experts.116", "model.encoder.language_model.layers.20.experts.117", "model.encoder.language_model.layers.20.experts.118", "model.encoder.language_model.layers.20.experts.119", "model.encoder.language_model.layers.20.experts.120", "model.encoder.language_model.layers.20.experts.121", "model.encoder.language_model.layers.20.experts.122", "model.encoder.language_model.layers.20.experts.123", "model.encoder.language_model.layers.20.experts.124", "model.encoder.language_model.layers.20.experts.125", "model.encoder.language_model.layers.20.experts.126", "model.encoder.language_model.layers.20.experts.127", "model.encoder.language_model.layers.21.router", "model.encoder.language_model.layers.21.router.proj", "model.encoder.language_model.layers.21.experts.0", "model.encoder.language_model.layers.21.experts.1", "model.encoder.language_model.layers.21.experts.2", "model.encoder.language_model.layers.21.experts.3", "model.encoder.language_model.layers.21.experts.4", "model.encoder.language_model.layers.21.experts.5", "model.encoder.language_model.layers.21.experts.6", "model.encoder.language_model.layers.21.experts.7", "model.encoder.language_model.layers.21.experts.8", "model.encoder.language_model.layers.21.experts.9", "model.encoder.language_model.layers.21.experts.10", "model.encoder.language_model.layers.21.experts.11", "model.encoder.language_model.layers.21.experts.12", "model.encoder.language_model.layers.21.experts.13", "model.encoder.language_model.layers.21.experts.14", "model.encoder.language_model.layers.21.experts.15", "model.encoder.language_model.layers.21.experts.16", "model.encoder.language_model.layers.21.experts.17", "model.encoder.language_model.layers.21.experts.18", "model.encoder.language_model.layers.21.experts.19", "model.encoder.language_model.layers.21.experts.20", "model.encoder.language_model.layers.21.experts.21", "model.encoder.language_model.layers.21.experts.22", "model.encoder.language_model.layers.21.experts.23", "model.encoder.language_model.layers.21.experts.24", "model.encoder.language_model.layers.21.experts.25", "model.encoder.language_model.layers.21.experts.26", "model.encoder.language_model.layers.21.experts.27", "model.encoder.language_model.layers.21.experts.28", "model.encoder.language_model.layers.21.experts.29", "model.encoder.language_model.layers.21.experts.30", "model.encoder.language_model.layers.21.experts.31", "model.encoder.language_model.layers.21.experts.32", "model.encoder.language_model.layers.21.experts.33", "model.encoder.language_model.layers.21.experts.34", "model.encoder.language_model.layers.21.experts.35", "model.encoder.language_model.layers.21.experts.36", "model.encoder.language_model.layers.21.experts.37", "model.encoder.language_model.layers.21.experts.38", "model.encoder.language_model.layers.21.experts.39", "model.encoder.language_model.layers.21.experts.40", "model.encoder.language_model.layers.21.experts.41", "model.encoder.language_model.layers.21.experts.42", "model.encoder.language_model.layers.21.experts.43", "model.encoder.language_model.layers.21.experts.44", "model.encoder.language_model.layers.21.experts.45", "model.encoder.language_model.layers.21.experts.46", "model.encoder.language_model.layers.21.experts.47", "model.encoder.language_model.layers.21.experts.48", "model.encoder.language_model.layers.21.experts.49", "model.encoder.language_model.layers.21.experts.50", "model.encoder.language_model.layers.21.experts.51", "model.encoder.language_model.layers.21.experts.52", "model.encoder.language_model.layers.21.experts.53", "model.encoder.language_model.layers.21.experts.54", "model.encoder.language_model.layers.21.experts.55", "model.encoder.language_model.layers.21.experts.56", "model.encoder.language_model.layers.21.experts.57", "model.encoder.language_model.layers.21.experts.58", "model.encoder.language_model.layers.21.experts.59", "model.encoder.language_model.layers.21.experts.60", "model.encoder.language_model.layers.21.experts.61", "model.encoder.language_model.layers.21.experts.62", "model.encoder.language_model.layers.21.experts.63", "model.encoder.language_model.layers.21.experts.64", "model.encoder.language_model.layers.21.experts.65", "model.encoder.language_model.layers.21.experts.66", "model.encoder.language_model.layers.21.experts.67", "model.encoder.language_model.layers.21.experts.68", "model.encoder.language_model.layers.21.experts.69", "model.encoder.language_model.layers.21.experts.70", "model.encoder.language_model.layers.21.experts.71", "model.encoder.language_model.layers.21.experts.72", "model.encoder.language_model.layers.21.experts.73", "model.encoder.language_model.layers.21.experts.74", "model.encoder.language_model.layers.21.experts.75", "model.encoder.language_model.layers.21.experts.76", "model.encoder.language_model.layers.21.experts.77", "model.encoder.language_model.layers.21.experts.78", "model.encoder.language_model.layers.21.experts.79", "model.encoder.language_model.layers.21.experts.80", "model.encoder.language_model.layers.21.experts.81", "model.encoder.language_model.layers.21.experts.82", "model.encoder.language_model.layers.21.experts.83", "model.encoder.language_model.layers.21.experts.84", "model.encoder.language_model.layers.21.experts.85", "model.encoder.language_model.layers.21.experts.86", "model.encoder.language_model.layers.21.experts.87", "model.encoder.language_model.layers.21.experts.88", "model.encoder.language_model.layers.21.experts.89", "model.encoder.language_model.layers.21.experts.90", "model.encoder.language_model.layers.21.experts.91", "model.encoder.language_model.layers.21.experts.92", "model.encoder.language_model.layers.21.experts.93", "model.encoder.language_model.layers.21.experts.94", "model.encoder.language_model.layers.21.experts.95", "model.encoder.language_model.layers.21.experts.96", "model.encoder.language_model.layers.21.experts.97", "model.encoder.language_model.layers.21.experts.98", "model.encoder.language_model.layers.21.experts.99", "model.encoder.language_model.layers.21.experts.100", "model.encoder.language_model.layers.21.experts.101", "model.encoder.language_model.layers.21.experts.102", "model.encoder.language_model.layers.21.experts.103", "model.encoder.language_model.layers.21.experts.104", "model.encoder.language_model.layers.21.experts.105", "model.encoder.language_model.layers.21.experts.106", "model.encoder.language_model.layers.21.experts.107", "model.encoder.language_model.layers.21.experts.108", "model.encoder.language_model.layers.21.experts.109", "model.encoder.language_model.layers.21.experts.110", "model.encoder.language_model.layers.21.experts.111", "model.encoder.language_model.layers.21.experts.112", "model.encoder.language_model.layers.21.experts.113", "model.encoder.language_model.layers.21.experts.114", "model.encoder.language_model.layers.21.experts.115", "model.encoder.language_model.layers.21.experts.116", "model.encoder.language_model.layers.21.experts.117", "model.encoder.language_model.layers.21.experts.118", "model.encoder.language_model.layers.21.experts.119", "model.encoder.language_model.layers.21.experts.120", "model.encoder.language_model.layers.21.experts.121", "model.encoder.language_model.layers.21.experts.122", "model.encoder.language_model.layers.21.experts.123", "model.encoder.language_model.layers.21.experts.124", "model.encoder.language_model.layers.21.experts.125", "model.encoder.language_model.layers.21.experts.126", "model.encoder.language_model.layers.21.experts.127", "model.encoder.language_model.layers.22.router", "model.encoder.language_model.layers.22.router.proj", "model.encoder.language_model.layers.22.experts.0", "model.encoder.language_model.layers.22.experts.1", "model.encoder.language_model.layers.22.experts.2", "model.encoder.language_model.layers.22.experts.3", "model.encoder.language_model.layers.22.experts.4", "model.encoder.language_model.layers.22.experts.5", "model.encoder.language_model.layers.22.experts.6", "model.encoder.language_model.layers.22.experts.7", "model.encoder.language_model.layers.22.experts.8", "model.encoder.language_model.layers.22.experts.9", "model.encoder.language_model.layers.22.experts.10", "model.encoder.language_model.layers.22.experts.11", "model.encoder.language_model.layers.22.experts.12", "model.encoder.language_model.layers.22.experts.13", "model.encoder.language_model.layers.22.experts.14", "model.encoder.language_model.layers.22.experts.15", "model.encoder.language_model.layers.22.experts.16", "model.encoder.language_model.layers.22.experts.17", "model.encoder.language_model.layers.22.experts.18", "model.encoder.language_model.layers.22.experts.19", "model.encoder.language_model.layers.22.experts.20", "model.encoder.language_model.layers.22.experts.21", "model.encoder.language_model.layers.22.experts.22", "model.encoder.language_model.layers.22.experts.23", "model.encoder.language_model.layers.22.experts.24", "model.encoder.language_model.layers.22.experts.25", "model.encoder.language_model.layers.22.experts.26", "model.encoder.language_model.layers.22.experts.27", "model.encoder.language_model.layers.22.experts.28", "model.encoder.language_model.layers.22.experts.29", "model.encoder.language_model.layers.22.experts.30", "model.encoder.language_model.layers.22.experts.31", "model.encoder.language_model.layers.22.experts.32", "model.encoder.language_model.layers.22.experts.33", "model.encoder.language_model.layers.22.experts.34", "model.encoder.language_model.layers.22.experts.35", "model.encoder.language_model.layers.22.experts.36", "model.encoder.language_model.layers.22.experts.37", "model.encoder.language_model.layers.22.experts.38", "model.encoder.language_model.layers.22.experts.39", "model.encoder.language_model.layers.22.experts.40", "model.encoder.language_model.layers.22.experts.41", "model.encoder.language_model.layers.22.experts.42", "model.encoder.language_model.layers.22.experts.43", "model.encoder.language_model.layers.22.experts.44", "model.encoder.language_model.layers.22.experts.45", "model.encoder.language_model.layers.22.experts.46", "model.encoder.language_model.layers.22.experts.47", "model.encoder.language_model.layers.22.experts.48", "model.encoder.language_model.layers.22.experts.49", "model.encoder.language_model.layers.22.experts.50", "model.encoder.language_model.layers.22.experts.51", "model.encoder.language_model.layers.22.experts.52", "model.encoder.language_model.layers.22.experts.53", "model.encoder.language_model.layers.22.experts.54", "model.encoder.language_model.layers.22.experts.55", "model.encoder.language_model.layers.22.experts.56", "model.encoder.language_model.layers.22.experts.57", "model.encoder.language_model.layers.22.experts.58", "model.encoder.language_model.layers.22.experts.59", "model.encoder.language_model.layers.22.experts.60", "model.encoder.language_model.layers.22.experts.61", "model.encoder.language_model.layers.22.experts.62", "model.encoder.language_model.layers.22.experts.63", "model.encoder.language_model.layers.22.experts.64", "model.encoder.language_model.layers.22.experts.65", "model.encoder.language_model.layers.22.experts.66", "model.encoder.language_model.layers.22.experts.67", "model.encoder.language_model.layers.22.experts.68", "model.encoder.language_model.layers.22.experts.69", "model.encoder.language_model.layers.22.experts.70", "model.encoder.language_model.layers.22.experts.71", "model.encoder.language_model.layers.22.experts.72", "model.encoder.language_model.layers.22.experts.73", "model.encoder.language_model.layers.22.experts.74", "model.encoder.language_model.layers.22.experts.75", "model.encoder.language_model.layers.22.experts.76", "model.encoder.language_model.layers.22.experts.77", "model.encoder.language_model.layers.22.experts.78", "model.encoder.language_model.layers.22.experts.79", "model.encoder.language_model.layers.22.experts.80", "model.encoder.language_model.layers.22.experts.81", "model.encoder.language_model.layers.22.experts.82", "model.encoder.language_model.layers.22.experts.83", "model.encoder.language_model.layers.22.experts.84", "model.encoder.language_model.layers.22.experts.85", "model.encoder.language_model.layers.22.experts.86", "model.encoder.language_model.layers.22.experts.87", "model.encoder.language_model.layers.22.experts.88", "model.encoder.language_model.layers.22.experts.89", "model.encoder.language_model.layers.22.experts.90", "model.encoder.language_model.layers.22.experts.91", "model.encoder.language_model.layers.22.experts.92", "model.encoder.language_model.layers.22.experts.93", "model.encoder.language_model.layers.22.experts.94", "model.encoder.language_model.layers.22.experts.95", "model.encoder.language_model.layers.22.experts.96", "model.encoder.language_model.layers.22.experts.97", "model.encoder.language_model.layers.22.experts.98", "model.encoder.language_model.layers.22.experts.99", "model.encoder.language_model.layers.22.experts.100", "model.encoder.language_model.layers.22.experts.101", "model.encoder.language_model.layers.22.experts.102", "model.encoder.language_model.layers.22.experts.103", "model.encoder.language_model.layers.22.experts.104", "model.encoder.language_model.layers.22.experts.105", "model.encoder.language_model.layers.22.experts.106", "model.encoder.language_model.layers.22.experts.107", "model.encoder.language_model.layers.22.experts.108", "model.encoder.language_model.layers.22.experts.109", "model.encoder.language_model.layers.22.experts.110", "model.encoder.language_model.layers.22.experts.111", "model.encoder.language_model.layers.22.experts.112", "model.encoder.language_model.layers.22.experts.113", "model.encoder.language_model.layers.22.experts.114", "model.encoder.language_model.layers.22.experts.115", "model.encoder.language_model.layers.22.experts.116", "model.encoder.language_model.layers.22.experts.117", "model.encoder.language_model.layers.22.experts.118", "model.encoder.language_model.layers.22.experts.119", "model.encoder.language_model.layers.22.experts.120", "model.encoder.language_model.layers.22.experts.121", "model.encoder.language_model.layers.22.experts.122", "model.encoder.language_model.layers.22.experts.123", "model.encoder.language_model.layers.22.experts.124", "model.encoder.language_model.layers.22.experts.125", "model.encoder.language_model.layers.22.experts.126", "model.encoder.language_model.layers.22.experts.127", "model.encoder.language_model.layers.23.router", "model.encoder.language_model.layers.23.router.proj", "model.encoder.language_model.layers.23.experts.0", "model.encoder.language_model.layers.23.experts.1", "model.encoder.language_model.layers.23.experts.2", "model.encoder.language_model.layers.23.experts.3", "model.encoder.language_model.layers.23.experts.4", "model.encoder.language_model.layers.23.experts.5", "model.encoder.language_model.layers.23.experts.6", "model.encoder.language_model.layers.23.experts.7", "model.encoder.language_model.layers.23.experts.8", "model.encoder.language_model.layers.23.experts.9", "model.encoder.language_model.layers.23.experts.10", "model.encoder.language_model.layers.23.experts.11", "model.encoder.language_model.layers.23.experts.12", "model.encoder.language_model.layers.23.experts.13", "model.encoder.language_model.layers.23.experts.14", "model.encoder.language_model.layers.23.experts.15", "model.encoder.language_model.layers.23.experts.16", "model.encoder.language_model.layers.23.experts.17", "model.encoder.language_model.layers.23.experts.18", "model.encoder.language_model.layers.23.experts.19", "model.encoder.language_model.layers.23.experts.20", "model.encoder.language_model.layers.23.experts.21", "model.encoder.language_model.layers.23.experts.22", "model.encoder.language_model.layers.23.experts.23", "model.encoder.language_model.layers.23.experts.24", "model.encoder.language_model.layers.23.experts.25", "model.encoder.language_model.layers.23.experts.26", "model.encoder.language_model.layers.23.experts.27", "model.encoder.language_model.layers.23.experts.28", "model.encoder.language_model.layers.23.experts.29", "model.encoder.language_model.layers.23.experts.30", "model.encoder.language_model.layers.23.experts.31", "model.encoder.language_model.layers.23.experts.32", "model.encoder.language_model.layers.23.experts.33", "model.encoder.language_model.layers.23.experts.34", "model.encoder.language_model.layers.23.experts.35", "model.encoder.language_model.layers.23.experts.36", "model.encoder.language_model.layers.23.experts.37", "model.encoder.language_model.layers.23.experts.38", "model.encoder.language_model.layers.23.experts.39", "model.encoder.language_model.layers.23.experts.40", "model.encoder.language_model.layers.23.experts.41", "model.encoder.language_model.layers.23.experts.42", "model.encoder.language_model.layers.23.experts.43", "model.encoder.language_model.layers.23.experts.44", "model.encoder.language_model.layers.23.experts.45", "model.encoder.language_model.layers.23.experts.46", "model.encoder.language_model.layers.23.experts.47", "model.encoder.language_model.layers.23.experts.48", "model.encoder.language_model.layers.23.experts.49", "model.encoder.language_model.layers.23.experts.50", "model.encoder.language_model.layers.23.experts.51", "model.encoder.language_model.layers.23.experts.52", "model.encoder.language_model.layers.23.experts.53", "model.encoder.language_model.layers.23.experts.54", "model.encoder.language_model.layers.23.experts.55", "model.encoder.language_model.layers.23.experts.56", "model.encoder.language_model.layers.23.experts.57", "model.encoder.language_model.layers.23.experts.58", "model.encoder.language_model.layers.23.experts.59", "model.encoder.language_model.layers.23.experts.60", "model.encoder.language_model.layers.23.experts.61", "model.encoder.language_model.layers.23.experts.62", "model.encoder.language_model.layers.23.experts.63", "model.encoder.language_model.layers.23.experts.64", "model.encoder.language_model.layers.23.experts.65", "model.encoder.language_model.layers.23.experts.66", "model.encoder.language_model.layers.23.experts.67", "model.encoder.language_model.layers.23.experts.68", "model.encoder.language_model.layers.23.experts.69", "model.encoder.language_model.layers.23.experts.70", "model.encoder.language_model.layers.23.experts.71", "model.encoder.language_model.layers.23.experts.72", "model.encoder.language_model.layers.23.experts.73", "model.encoder.language_model.layers.23.experts.74", "model.encoder.language_model.layers.23.experts.75", "model.encoder.language_model.layers.23.experts.76", "model.encoder.language_model.layers.23.experts.77", "model.encoder.language_model.layers.23.experts.78", "model.encoder.language_model.layers.23.experts.79", "model.encoder.language_model.layers.23.experts.80", "model.encoder.language_model.layers.23.experts.81", "model.encoder.language_model.layers.23.experts.82", "model.encoder.language_model.layers.23.experts.83", "model.encoder.language_model.layers.23.experts.84", "model.encoder.language_model.layers.23.experts.85", "model.encoder.language_model.layers.23.experts.86", "model.encoder.language_model.layers.23.experts.87", "model.encoder.language_model.layers.23.experts.88", "model.encoder.language_model.layers.23.experts.89", "model.encoder.language_model.layers.23.experts.90", "model.encoder.language_model.layers.23.experts.91", "model.encoder.language_model.layers.23.experts.92", "model.encoder.language_model.layers.23.experts.93", "model.encoder.language_model.layers.23.experts.94", "model.encoder.language_model.layers.23.experts.95", "model.encoder.language_model.layers.23.experts.96", "model.encoder.language_model.layers.23.experts.97", "model.encoder.language_model.layers.23.experts.98", "model.encoder.language_model.layers.23.experts.99", "model.encoder.language_model.layers.23.experts.100", "model.encoder.language_model.layers.23.experts.101", "model.encoder.language_model.layers.23.experts.102", "model.encoder.language_model.layers.23.experts.103", "model.encoder.language_model.layers.23.experts.104", "model.encoder.language_model.layers.23.experts.105", "model.encoder.language_model.layers.23.experts.106", "model.encoder.language_model.layers.23.experts.107", "model.encoder.language_model.layers.23.experts.108", "model.encoder.language_model.layers.23.experts.109", "model.encoder.language_model.layers.23.experts.110", "model.encoder.language_model.layers.23.experts.111", "model.encoder.language_model.layers.23.experts.112", "model.encoder.language_model.layers.23.experts.113", "model.encoder.language_model.layers.23.experts.114", "model.encoder.language_model.layers.23.experts.115", "model.encoder.language_model.layers.23.experts.116", "model.encoder.language_model.layers.23.experts.117", "model.encoder.language_model.layers.23.experts.118", "model.encoder.language_model.layers.23.experts.119", "model.encoder.language_model.layers.23.experts.120", "model.encoder.language_model.layers.23.experts.121", "model.encoder.language_model.layers.23.experts.122", "model.encoder.language_model.layers.23.experts.123", "model.encoder.language_model.layers.23.experts.124", "model.encoder.language_model.layers.23.experts.125", "model.encoder.language_model.layers.23.experts.126", "model.encoder.language_model.layers.23.experts.127", "model.encoder.language_model.layers.24.router", "model.encoder.language_model.layers.24.router.proj", "model.encoder.language_model.layers.24.experts.0", "model.encoder.language_model.layers.24.experts.1", "model.encoder.language_model.layers.24.experts.2", "model.encoder.language_model.layers.24.experts.3", "model.encoder.language_model.layers.24.experts.4", "model.encoder.language_model.layers.24.experts.5", "model.encoder.language_model.layers.24.experts.6", "model.encoder.language_model.layers.24.experts.7", "model.encoder.language_model.layers.24.experts.8", "model.encoder.language_model.layers.24.experts.9", "model.encoder.language_model.layers.24.experts.10", "model.encoder.language_model.layers.24.experts.11", "model.encoder.language_model.layers.24.experts.12", "model.encoder.language_model.layers.24.experts.13", "model.encoder.language_model.layers.24.experts.14", "model.encoder.language_model.layers.24.experts.15", "model.encoder.language_model.layers.24.experts.16", "model.encoder.language_model.layers.24.experts.17", "model.encoder.language_model.layers.24.experts.18", "model.encoder.language_model.layers.24.experts.19", "model.encoder.language_model.layers.24.experts.20", "model.encoder.language_model.layers.24.experts.21", "model.encoder.language_model.layers.24.experts.22", "model.encoder.language_model.layers.24.experts.23", "model.encoder.language_model.layers.24.experts.24", "model.encoder.language_model.layers.24.experts.25", "model.encoder.language_model.layers.24.experts.26", "model.encoder.language_model.layers.24.experts.27", "model.encoder.language_model.layers.24.experts.28", "model.encoder.language_model.layers.24.experts.29", "model.encoder.language_model.layers.24.experts.30", "model.encoder.language_model.layers.24.experts.31", "model.encoder.language_model.layers.24.experts.32", "model.encoder.language_model.layers.24.experts.33", "model.encoder.language_model.layers.24.experts.34", "model.encoder.language_model.layers.24.experts.35", "model.encoder.language_model.layers.24.experts.36", "model.encoder.language_model.layers.24.experts.37", "model.encoder.language_model.layers.24.experts.38", "model.encoder.language_model.layers.24.experts.39", "model.encoder.language_model.layers.24.experts.40", "model.encoder.language_model.layers.24.experts.41", "model.encoder.language_model.layers.24.experts.42", "model.encoder.language_model.layers.24.experts.43", "model.encoder.language_model.layers.24.experts.44", "model.encoder.language_model.layers.24.experts.45", "model.encoder.language_model.layers.24.experts.46", "model.encoder.language_model.layers.24.experts.47", "model.encoder.language_model.layers.24.experts.48", "model.encoder.language_model.layers.24.experts.49", "model.encoder.language_model.layers.24.experts.50", "model.encoder.language_model.layers.24.experts.51", "model.encoder.language_model.layers.24.experts.52", "model.encoder.language_model.layers.24.experts.53", "model.encoder.language_model.layers.24.experts.54", "model.encoder.language_model.layers.24.experts.55", "model.encoder.language_model.layers.24.experts.56", "model.encoder.language_model.layers.24.experts.57", "model.encoder.language_model.layers.24.experts.58", "model.encoder.language_model.layers.24.experts.59", "model.encoder.language_model.layers.24.experts.60", "model.encoder.language_model.layers.24.experts.61", "model.encoder.language_model.layers.24.experts.62", "model.encoder.language_model.layers.24.experts.63", "model.encoder.language_model.layers.24.experts.64", "model.encoder.language_model.layers.24.experts.65", "model.encoder.language_model.layers.24.experts.66", "model.encoder.language_model.layers.24.experts.67", "model.encoder.language_model.layers.24.experts.68", "model.encoder.language_model.layers.24.experts.69", "model.encoder.language_model.layers.24.experts.70", "model.encoder.language_model.layers.24.experts.71", "model.encoder.language_model.layers.24.experts.72", "model.encoder.language_model.layers.24.experts.73", "model.encoder.language_model.layers.24.experts.74", "model.encoder.language_model.layers.24.experts.75", "model.encoder.language_model.layers.24.experts.76", "model.encoder.language_model.layers.24.experts.77", "model.encoder.language_model.layers.24.experts.78", "model.encoder.language_model.layers.24.experts.79", "model.encoder.language_model.layers.24.experts.80", "model.encoder.language_model.layers.24.experts.81", "model.encoder.language_model.layers.24.experts.82", "model.encoder.language_model.layers.24.experts.83", "model.encoder.language_model.layers.24.experts.84", "model.encoder.language_model.layers.24.experts.85", "model.encoder.language_model.layers.24.experts.86", "model.encoder.language_model.layers.24.experts.87", "model.encoder.language_model.layers.24.experts.88", "model.encoder.language_model.layers.24.experts.89", "model.encoder.language_model.layers.24.experts.90", "model.encoder.language_model.layers.24.experts.91", "model.encoder.language_model.layers.24.experts.92", "model.encoder.language_model.layers.24.experts.93", "model.encoder.language_model.layers.24.experts.94", "model.encoder.language_model.layers.24.experts.95", "model.encoder.language_model.layers.24.experts.96", "model.encoder.language_model.layers.24.experts.97", "model.encoder.language_model.layers.24.experts.98", "model.encoder.language_model.layers.24.experts.99", "model.encoder.language_model.layers.24.experts.100", "model.encoder.language_model.layers.24.experts.101", "model.encoder.language_model.layers.24.experts.102", "model.encoder.language_model.layers.24.experts.103", "model.encoder.language_model.layers.24.experts.104", "model.encoder.language_model.layers.24.experts.105", "model.encoder.language_model.layers.24.experts.106", "model.encoder.language_model.layers.24.experts.107", "model.encoder.language_model.layers.24.experts.108", "model.encoder.language_model.layers.24.experts.109", "model.encoder.language_model.layers.24.experts.110", "model.encoder.language_model.layers.24.experts.111", "model.encoder.language_model.layers.24.experts.112", "model.encoder.language_model.layers.24.experts.113", "model.encoder.language_model.layers.24.experts.114", "model.encoder.language_model.layers.24.experts.115", "model.encoder.language_model.layers.24.experts.116", "model.encoder.language_model.layers.24.experts.117", "model.encoder.language_model.layers.24.experts.118", "model.encoder.language_model.layers.24.experts.119", "model.encoder.language_model.layers.24.experts.120", "model.encoder.language_model.layers.24.experts.121", "model.encoder.language_model.layers.24.experts.122", "model.encoder.language_model.layers.24.experts.123", "model.encoder.language_model.layers.24.experts.124", "model.encoder.language_model.layers.24.experts.125", "model.encoder.language_model.layers.24.experts.126", "model.encoder.language_model.layers.24.experts.127", "model.encoder.language_model.layers.25.router", "model.encoder.language_model.layers.25.router.proj", "model.encoder.language_model.layers.25.experts.0", "model.encoder.language_model.layers.25.experts.1", "model.encoder.language_model.layers.25.experts.2", "model.encoder.language_model.layers.25.experts.3", "model.encoder.language_model.layers.25.experts.4", "model.encoder.language_model.layers.25.experts.5", "model.encoder.language_model.layers.25.experts.6", "model.encoder.language_model.layers.25.experts.7", "model.encoder.language_model.layers.25.experts.8", "model.encoder.language_model.layers.25.experts.9", "model.encoder.language_model.layers.25.experts.10", "model.encoder.language_model.layers.25.experts.11", "model.encoder.language_model.layers.25.experts.12", "model.encoder.language_model.layers.25.experts.13", "model.encoder.language_model.layers.25.experts.14", "model.encoder.language_model.layers.25.experts.15", "model.encoder.language_model.layers.25.experts.16", "model.encoder.language_model.layers.25.experts.17", "model.encoder.language_model.layers.25.experts.18", "model.encoder.language_model.layers.25.experts.19", "model.encoder.language_model.layers.25.experts.20", "model.encoder.language_model.layers.25.experts.21", "model.encoder.language_model.layers.25.experts.22", "model.encoder.language_model.layers.25.experts.23", "model.encoder.language_model.layers.25.experts.24", "model.encoder.language_model.layers.25.experts.25", "model.encoder.language_model.layers.25.experts.26", "model.encoder.language_model.layers.25.experts.27", "model.encoder.language_model.layers.25.experts.28", "model.encoder.language_model.layers.25.experts.29", "model.encoder.language_model.layers.25.experts.30", "model.encoder.language_model.layers.25.experts.31", "model.encoder.language_model.layers.25.experts.32", "model.encoder.language_model.layers.25.experts.33", "model.encoder.language_model.layers.25.experts.34", "model.encoder.language_model.layers.25.experts.35", "model.encoder.language_model.layers.25.experts.36", "model.encoder.language_model.layers.25.experts.37", "model.encoder.language_model.layers.25.experts.38", "model.encoder.language_model.layers.25.experts.39", "model.encoder.language_model.layers.25.experts.40", "model.encoder.language_model.layers.25.experts.41", "model.encoder.language_model.layers.25.experts.42", "model.encoder.language_model.layers.25.experts.43", "model.encoder.language_model.layers.25.experts.44", "model.encoder.language_model.layers.25.experts.45", "model.encoder.language_model.layers.25.experts.46", "model.encoder.language_model.layers.25.experts.47", "model.encoder.language_model.layers.25.experts.48", "model.encoder.language_model.layers.25.experts.49", "model.encoder.language_model.layers.25.experts.50", "model.encoder.language_model.layers.25.experts.51", "model.encoder.language_model.layers.25.experts.52", "model.encoder.language_model.layers.25.experts.53", "model.encoder.language_model.layers.25.experts.54", "model.encoder.language_model.layers.25.experts.55", "model.encoder.language_model.layers.25.experts.56", "model.encoder.language_model.layers.25.experts.57", "model.encoder.language_model.layers.25.experts.58", "model.encoder.language_model.layers.25.experts.59", "model.encoder.language_model.layers.25.experts.60", "model.encoder.language_model.layers.25.experts.61", "model.encoder.language_model.layers.25.experts.62", "model.encoder.language_model.layers.25.experts.63", "model.encoder.language_model.layers.25.experts.64", "model.encoder.language_model.layers.25.experts.65", "model.encoder.language_model.layers.25.experts.66", "model.encoder.language_model.layers.25.experts.67", "model.encoder.language_model.layers.25.experts.68", "model.encoder.language_model.layers.25.experts.69", "model.encoder.language_model.layers.25.experts.70", "model.encoder.language_model.layers.25.experts.71", "model.encoder.language_model.layers.25.experts.72", "model.encoder.language_model.layers.25.experts.73", "model.encoder.language_model.layers.25.experts.74", "model.encoder.language_model.layers.25.experts.75", "model.encoder.language_model.layers.25.experts.76", "model.encoder.language_model.layers.25.experts.77", "model.encoder.language_model.layers.25.experts.78", "model.encoder.language_model.layers.25.experts.79", "model.encoder.language_model.layers.25.experts.80", "model.encoder.language_model.layers.25.experts.81", "model.encoder.language_model.layers.25.experts.82", "model.encoder.language_model.layers.25.experts.83", "model.encoder.language_model.layers.25.experts.84", "model.encoder.language_model.layers.25.experts.85", "model.encoder.language_model.layers.25.experts.86", "model.encoder.language_model.layers.25.experts.87", "model.encoder.language_model.layers.25.experts.88", "model.encoder.language_model.layers.25.experts.89", "model.encoder.language_model.layers.25.experts.90", "model.encoder.language_model.layers.25.experts.91", "model.encoder.language_model.layers.25.experts.92", "model.encoder.language_model.layers.25.experts.93", "model.encoder.language_model.layers.25.experts.94", "model.encoder.language_model.layers.25.experts.95", "model.encoder.language_model.layers.25.experts.96", "model.encoder.language_model.layers.25.experts.97", "model.encoder.language_model.layers.25.experts.98", "model.encoder.language_model.layers.25.experts.99", "model.encoder.language_model.layers.25.experts.100", "model.encoder.language_model.layers.25.experts.101", "model.encoder.language_model.layers.25.experts.102", "model.encoder.language_model.layers.25.experts.103", "model.encoder.language_model.layers.25.experts.104", "model.encoder.language_model.layers.25.experts.105", "model.encoder.language_model.layers.25.experts.106", "model.encoder.language_model.layers.25.experts.107", "model.encoder.language_model.layers.25.experts.108", "model.encoder.language_model.layers.25.experts.109", "model.encoder.language_model.layers.25.experts.110", "model.encoder.language_model.layers.25.experts.111", "model.encoder.language_model.layers.25.experts.112", "model.encoder.language_model.layers.25.experts.113", "model.encoder.language_model.layers.25.experts.114", "model.encoder.language_model.layers.25.experts.115", "model.encoder.language_model.layers.25.experts.116", "model.encoder.language_model.layers.25.experts.117", "model.encoder.language_model.layers.25.experts.118", "model.encoder.language_model.layers.25.experts.119", "model.encoder.language_model.layers.25.experts.120", "model.encoder.language_model.layers.25.experts.121", "model.encoder.language_model.layers.25.experts.122", "model.encoder.language_model.layers.25.experts.123", "model.encoder.language_model.layers.25.experts.124", "model.encoder.language_model.layers.25.experts.125", "model.encoder.language_model.layers.25.experts.126", "model.encoder.language_model.layers.25.experts.127", "model.encoder.language_model.layers.26.router", "model.encoder.language_model.layers.26.router.proj", "model.encoder.language_model.layers.26.experts.0", "model.encoder.language_model.layers.26.experts.1", "model.encoder.language_model.layers.26.experts.2", "model.encoder.language_model.layers.26.experts.3", "model.encoder.language_model.layers.26.experts.4", "model.encoder.language_model.layers.26.experts.5", "model.encoder.language_model.layers.26.experts.6", "model.encoder.language_model.layers.26.experts.7", "model.encoder.language_model.layers.26.experts.8", "model.encoder.language_model.layers.26.experts.9", "model.encoder.language_model.layers.26.experts.10", "model.encoder.language_model.layers.26.experts.11", "model.encoder.language_model.layers.26.experts.12", "model.encoder.language_model.layers.26.experts.13", "model.encoder.language_model.layers.26.experts.14", "model.encoder.language_model.layers.26.experts.15", "model.encoder.language_model.layers.26.experts.16", "model.encoder.language_model.layers.26.experts.17", "model.encoder.language_model.layers.26.experts.18", "model.encoder.language_model.layers.26.experts.19", "model.encoder.language_model.layers.26.experts.20", "model.encoder.language_model.layers.26.experts.21", "model.encoder.language_model.layers.26.experts.22", "model.encoder.language_model.layers.26.experts.23", "model.encoder.language_model.layers.26.experts.24", "model.encoder.language_model.layers.26.experts.25", "model.encoder.language_model.layers.26.experts.26", "model.encoder.language_model.layers.26.experts.27", "model.encoder.language_model.layers.26.experts.28", "model.encoder.language_model.layers.26.experts.29", "model.encoder.language_model.layers.26.experts.30", "model.encoder.language_model.layers.26.experts.31", "model.encoder.language_model.layers.26.experts.32", "model.encoder.language_model.layers.26.experts.33", "model.encoder.language_model.layers.26.experts.34", "model.encoder.language_model.layers.26.experts.35", "model.encoder.language_model.layers.26.experts.36", "model.encoder.language_model.layers.26.experts.37", "model.encoder.language_model.layers.26.experts.38", "model.encoder.language_model.layers.26.experts.39", "model.encoder.language_model.layers.26.experts.40", "model.encoder.language_model.layers.26.experts.41", "model.encoder.language_model.layers.26.experts.42", "model.encoder.language_model.layers.26.experts.43", "model.encoder.language_model.layers.26.experts.44", "model.encoder.language_model.layers.26.experts.45", "model.encoder.language_model.layers.26.experts.46", "model.encoder.language_model.layers.26.experts.47", "model.encoder.language_model.layers.26.experts.48", "model.encoder.language_model.layers.26.experts.49", "model.encoder.language_model.layers.26.experts.50", "model.encoder.language_model.layers.26.experts.51", "model.encoder.language_model.layers.26.experts.52", "model.encoder.language_model.layers.26.experts.53", "model.encoder.language_model.layers.26.experts.54", "model.encoder.language_model.layers.26.experts.55", "model.encoder.language_model.layers.26.experts.56", "model.encoder.language_model.layers.26.experts.57", "model.encoder.language_model.layers.26.experts.58", "model.encoder.language_model.layers.26.experts.59", "model.encoder.language_model.layers.26.experts.60", "model.encoder.language_model.layers.26.experts.61", "model.encoder.language_model.layers.26.experts.62", "model.encoder.language_model.layers.26.experts.63", "model.encoder.language_model.layers.26.experts.64", "model.encoder.language_model.layers.26.experts.65", "model.encoder.language_model.layers.26.experts.66", "model.encoder.language_model.layers.26.experts.67", "model.encoder.language_model.layers.26.experts.68", "model.encoder.language_model.layers.26.experts.69", "model.encoder.language_model.layers.26.experts.70", "model.encoder.language_model.layers.26.experts.71", "model.encoder.language_model.layers.26.experts.72", "model.encoder.language_model.layers.26.experts.73", "model.encoder.language_model.layers.26.experts.74", "model.encoder.language_model.layers.26.experts.75", "model.encoder.language_model.layers.26.experts.76", "model.encoder.language_model.layers.26.experts.77", "model.encoder.language_model.layers.26.experts.78", "model.encoder.language_model.layers.26.experts.79", "model.encoder.language_model.layers.26.experts.80", "model.encoder.language_model.layers.26.experts.81", "model.encoder.language_model.layers.26.experts.82", "model.encoder.language_model.layers.26.experts.83", "model.encoder.language_model.layers.26.experts.84", "model.encoder.language_model.layers.26.experts.85", "model.encoder.language_model.layers.26.experts.86", "model.encoder.language_model.layers.26.experts.87", "model.encoder.language_model.layers.26.experts.88", "model.encoder.language_model.layers.26.experts.89", "model.encoder.language_model.layers.26.experts.90", "model.encoder.language_model.layers.26.experts.91", "model.encoder.language_model.layers.26.experts.92", "model.encoder.language_model.layers.26.experts.93", "model.encoder.language_model.layers.26.experts.94", "model.encoder.language_model.layers.26.experts.95", "model.encoder.language_model.layers.26.experts.96", "model.encoder.language_model.layers.26.experts.97", "model.encoder.language_model.layers.26.experts.98", "model.encoder.language_model.layers.26.experts.99", "model.encoder.language_model.layers.26.experts.100", "model.encoder.language_model.layers.26.experts.101", "model.encoder.language_model.layers.26.experts.102", "model.encoder.language_model.layers.26.experts.103", "model.encoder.language_model.layers.26.experts.104", "model.encoder.language_model.layers.26.experts.105", "model.encoder.language_model.layers.26.experts.106", "model.encoder.language_model.layers.26.experts.107", "model.encoder.language_model.layers.26.experts.108", "model.encoder.language_model.layers.26.experts.109", "model.encoder.language_model.layers.26.experts.110", "model.encoder.language_model.layers.26.experts.111", "model.encoder.language_model.layers.26.experts.112", "model.encoder.language_model.layers.26.experts.113", "model.encoder.language_model.layers.26.experts.114", "model.encoder.language_model.layers.26.experts.115", "model.encoder.language_model.layers.26.experts.116", "model.encoder.language_model.layers.26.experts.117", "model.encoder.language_model.layers.26.experts.118", "model.encoder.language_model.layers.26.experts.119", "model.encoder.language_model.layers.26.experts.120", "model.encoder.language_model.layers.26.experts.121", "model.encoder.language_model.layers.26.experts.122", "model.encoder.language_model.layers.26.experts.123", "model.encoder.language_model.layers.26.experts.124", "model.encoder.language_model.layers.26.experts.125", "model.encoder.language_model.layers.26.experts.126", "model.encoder.language_model.layers.26.experts.127", "model.encoder.language_model.layers.27.router", "model.encoder.language_model.layers.27.router.proj", "model.encoder.language_model.layers.27.experts.0", "model.encoder.language_model.layers.27.experts.1", "model.encoder.language_model.layers.27.experts.2", "model.encoder.language_model.layers.27.experts.3", "model.encoder.language_model.layers.27.experts.4", "model.encoder.language_model.layers.27.experts.5", "model.encoder.language_model.layers.27.experts.6", "model.encoder.language_model.layers.27.experts.7", "model.encoder.language_model.layers.27.experts.8", "model.encoder.language_model.layers.27.experts.9", "model.encoder.language_model.layers.27.experts.10", "model.encoder.language_model.layers.27.experts.11", "model.encoder.language_model.layers.27.experts.12", "model.encoder.language_model.layers.27.experts.13", "model.encoder.language_model.layers.27.experts.14", "model.encoder.language_model.layers.27.experts.15", "model.encoder.language_model.layers.27.experts.16", "model.encoder.language_model.layers.27.experts.17", "model.encoder.language_model.layers.27.experts.18", "model.encoder.language_model.layers.27.experts.19", "model.encoder.language_model.layers.27.experts.20", "model.encoder.language_model.layers.27.experts.21", "model.encoder.language_model.layers.27.experts.22", "model.encoder.language_model.layers.27.experts.23", "model.encoder.language_model.layers.27.experts.24", "model.encoder.language_model.layers.27.experts.25", "model.encoder.language_model.layers.27.experts.26", "model.encoder.language_model.layers.27.experts.27", "model.encoder.language_model.layers.27.experts.28", "model.encoder.language_model.layers.27.experts.29", "model.encoder.language_model.layers.27.experts.30", "model.encoder.language_model.layers.27.experts.31", "model.encoder.language_model.layers.27.experts.32", "model.encoder.language_model.layers.27.experts.33", "model.encoder.language_model.layers.27.experts.34", "model.encoder.language_model.layers.27.experts.35", "model.encoder.language_model.layers.27.experts.36", "model.encoder.language_model.layers.27.experts.37", "model.encoder.language_model.layers.27.experts.38", "model.encoder.language_model.layers.27.experts.39", "model.encoder.language_model.layers.27.experts.40", "model.encoder.language_model.layers.27.experts.41", "model.encoder.language_model.layers.27.experts.42", "model.encoder.language_model.layers.27.experts.43", "model.encoder.language_model.layers.27.experts.44", "model.encoder.language_model.layers.27.experts.45", "model.encoder.language_model.layers.27.experts.46", "model.encoder.language_model.layers.27.experts.47", "model.encoder.language_model.layers.27.experts.48", "model.encoder.language_model.layers.27.experts.49", "model.encoder.language_model.layers.27.experts.50", "model.encoder.language_model.layers.27.experts.51", "model.encoder.language_model.layers.27.experts.52", "model.encoder.language_model.layers.27.experts.53", "model.encoder.language_model.layers.27.experts.54", "model.encoder.language_model.layers.27.experts.55", "model.encoder.language_model.layers.27.experts.56", "model.encoder.language_model.layers.27.experts.57", "model.encoder.language_model.layers.27.experts.58", "model.encoder.language_model.layers.27.experts.59", "model.encoder.language_model.layers.27.experts.60", "model.encoder.language_model.layers.27.experts.61", "model.encoder.language_model.layers.27.experts.62", "model.encoder.language_model.layers.27.experts.63", "model.encoder.language_model.layers.27.experts.64", "model.encoder.language_model.layers.27.experts.65", "model.encoder.language_model.layers.27.experts.66", "model.encoder.language_model.layers.27.experts.67", "model.encoder.language_model.layers.27.experts.68", "model.encoder.language_model.layers.27.experts.69", "model.encoder.language_model.layers.27.experts.70", "model.encoder.language_model.layers.27.experts.71", "model.encoder.language_model.layers.27.experts.72", "model.encoder.language_model.layers.27.experts.73", "model.encoder.language_model.layers.27.experts.74", "model.encoder.language_model.layers.27.experts.75", "model.encoder.language_model.layers.27.experts.76", "model.encoder.language_model.layers.27.experts.77", "model.encoder.language_model.layers.27.experts.78", "model.encoder.language_model.layers.27.experts.79", "model.encoder.language_model.layers.27.experts.80", "model.encoder.language_model.layers.27.experts.81", "model.encoder.language_model.layers.27.experts.82", "model.encoder.language_model.layers.27.experts.83", "model.encoder.language_model.layers.27.experts.84", "model.encoder.language_model.layers.27.experts.85", "model.encoder.language_model.layers.27.experts.86", "model.encoder.language_model.layers.27.experts.87", "model.encoder.language_model.layers.27.experts.88", "model.encoder.language_model.layers.27.experts.89", "model.encoder.language_model.layers.27.experts.90", "model.encoder.language_model.layers.27.experts.91", "model.encoder.language_model.layers.27.experts.92", "model.encoder.language_model.layers.27.experts.93", "model.encoder.language_model.layers.27.experts.94", "model.encoder.language_model.layers.27.experts.95", "model.encoder.language_model.layers.27.experts.96", "model.encoder.language_model.layers.27.experts.97", "model.encoder.language_model.layers.27.experts.98", "model.encoder.language_model.layers.27.experts.99", "model.encoder.language_model.layers.27.experts.100", "model.encoder.language_model.layers.27.experts.101", "model.encoder.language_model.layers.27.experts.102", "model.encoder.language_model.layers.27.experts.103", "model.encoder.language_model.layers.27.experts.104", "model.encoder.language_model.layers.27.experts.105", "model.encoder.language_model.layers.27.experts.106", "model.encoder.language_model.layers.27.experts.107", "model.encoder.language_model.layers.27.experts.108", "model.encoder.language_model.layers.27.experts.109", "model.encoder.language_model.layers.27.experts.110", "model.encoder.language_model.layers.27.experts.111", "model.encoder.language_model.layers.27.experts.112", "model.encoder.language_model.layers.27.experts.113", "model.encoder.language_model.layers.27.experts.114", "model.encoder.language_model.layers.27.experts.115", "model.encoder.language_model.layers.27.experts.116", "model.encoder.language_model.layers.27.experts.117", "model.encoder.language_model.layers.27.experts.118", "model.encoder.language_model.layers.27.experts.119", "model.encoder.language_model.layers.27.experts.120", "model.encoder.language_model.layers.27.experts.121", "model.encoder.language_model.layers.27.experts.122", "model.encoder.language_model.layers.27.experts.123", "model.encoder.language_model.layers.27.experts.124", "model.encoder.language_model.layers.27.experts.125", "model.encoder.language_model.layers.27.experts.126", "model.encoder.language_model.layers.27.experts.127", "model.encoder.language_model.layers.28.router", "model.encoder.language_model.layers.28.router.proj", "model.encoder.language_model.layers.28.experts.0", "model.encoder.language_model.layers.28.experts.1", "model.encoder.language_model.layers.28.experts.2", "model.encoder.language_model.layers.28.experts.3", "model.encoder.language_model.layers.28.experts.4", "model.encoder.language_model.layers.28.experts.5", "model.encoder.language_model.layers.28.experts.6", "model.encoder.language_model.layers.28.experts.7", "model.encoder.language_model.layers.28.experts.8", "model.encoder.language_model.layers.28.experts.9", "model.encoder.language_model.layers.28.experts.10", "model.encoder.language_model.layers.28.experts.11", "model.encoder.language_model.layers.28.experts.12", "model.encoder.language_model.layers.28.experts.13", "model.encoder.language_model.layers.28.experts.14", "model.encoder.language_model.layers.28.experts.15", "model.encoder.language_model.layers.28.experts.16", "model.encoder.language_model.layers.28.experts.17", "model.encoder.language_model.layers.28.experts.18", "model.encoder.language_model.layers.28.experts.19", "model.encoder.language_model.layers.28.experts.20", "model.encoder.language_model.layers.28.experts.21", "model.encoder.language_model.layers.28.experts.22", "model.encoder.language_model.layers.28.experts.23", "model.encoder.language_model.layers.28.experts.24", "model.encoder.language_model.layers.28.experts.25", "model.encoder.language_model.layers.28.experts.26", "model.encoder.language_model.layers.28.experts.27", "model.encoder.language_model.layers.28.experts.28", "model.encoder.language_model.layers.28.experts.29", "model.encoder.language_model.layers.28.experts.30", "model.encoder.language_model.layers.28.experts.31", "model.encoder.language_model.layers.28.experts.32", "model.encoder.language_model.layers.28.experts.33", "model.encoder.language_model.layers.28.experts.34", "model.encoder.language_model.layers.28.experts.35", "model.encoder.language_model.layers.28.experts.36", "model.encoder.language_model.layers.28.experts.37", "model.encoder.language_model.layers.28.experts.38", "model.encoder.language_model.layers.28.experts.39", "model.encoder.language_model.layers.28.experts.40", "model.encoder.language_model.layers.28.experts.41", "model.encoder.language_model.layers.28.experts.42", "model.encoder.language_model.layers.28.experts.43", "model.encoder.language_model.layers.28.experts.44", "model.encoder.language_model.layers.28.experts.45", "model.encoder.language_model.layers.28.experts.46", "model.encoder.language_model.layers.28.experts.47", "model.encoder.language_model.layers.28.experts.48", "model.encoder.language_model.layers.28.experts.49", "model.encoder.language_model.layers.28.experts.50", "model.encoder.language_model.layers.28.experts.51", "model.encoder.language_model.layers.28.experts.52", "model.encoder.language_model.layers.28.experts.53", "model.encoder.language_model.layers.28.experts.54", "model.encoder.language_model.layers.28.experts.55", "model.encoder.language_model.layers.28.experts.56", "model.encoder.language_model.layers.28.experts.57", "model.encoder.language_model.layers.28.experts.58", "model.encoder.language_model.layers.28.experts.59", "model.encoder.language_model.layers.28.experts.60", "model.encoder.language_model.layers.28.experts.61", "model.encoder.language_model.layers.28.experts.62", "model.encoder.language_model.layers.28.experts.63", "model.encoder.language_model.layers.28.experts.64", "model.encoder.language_model.layers.28.experts.65", "model.encoder.language_model.layers.28.experts.66", "model.encoder.language_model.layers.28.experts.67", "model.encoder.language_model.layers.28.experts.68", "model.encoder.language_model.layers.28.experts.69", "model.encoder.language_model.layers.28.experts.70", "model.encoder.language_model.layers.28.experts.71", "model.encoder.language_model.layers.28.experts.72", "model.encoder.language_model.layers.28.experts.73", "model.encoder.language_model.layers.28.experts.74", "model.encoder.language_model.layers.28.experts.75", "model.encoder.language_model.layers.28.experts.76", "model.encoder.language_model.layers.28.experts.77", "model.encoder.language_model.layers.28.experts.78", "model.encoder.language_model.layers.28.experts.79", "model.encoder.language_model.layers.28.experts.80", "model.encoder.language_model.layers.28.experts.81", "model.encoder.language_model.layers.28.experts.82", "model.encoder.language_model.layers.28.experts.83", "model.encoder.language_model.layers.28.experts.84", "model.encoder.language_model.layers.28.experts.85", "model.encoder.language_model.layers.28.experts.86", "model.encoder.language_model.layers.28.experts.87", "model.encoder.language_model.layers.28.experts.88", "model.encoder.language_model.layers.28.experts.89", "model.encoder.language_model.layers.28.experts.90", "model.encoder.language_model.layers.28.experts.91", "model.encoder.language_model.layers.28.experts.92", "model.encoder.language_model.layers.28.experts.93", "model.encoder.language_model.layers.28.experts.94", "model.encoder.language_model.layers.28.experts.95", "model.encoder.language_model.layers.28.experts.96", "model.encoder.language_model.layers.28.experts.97", "model.encoder.language_model.layers.28.experts.98", "model.encoder.language_model.layers.28.experts.99", "model.encoder.language_model.layers.28.experts.100", "model.encoder.language_model.layers.28.experts.101", "model.encoder.language_model.layers.28.experts.102", "model.encoder.language_model.layers.28.experts.103", "model.encoder.language_model.layers.28.experts.104", "model.encoder.language_model.layers.28.experts.105", "model.encoder.language_model.layers.28.experts.106", "model.encoder.language_model.layers.28.experts.107", "model.encoder.language_model.layers.28.experts.108", "model.encoder.language_model.layers.28.experts.109", "model.encoder.language_model.layers.28.experts.110", "model.encoder.language_model.layers.28.experts.111", "model.encoder.language_model.layers.28.experts.112", "model.encoder.language_model.layers.28.experts.113", "model.encoder.language_model.layers.28.experts.114", "model.encoder.language_model.layers.28.experts.115", "model.encoder.language_model.layers.28.experts.116", "model.encoder.language_model.layers.28.experts.117", "model.encoder.language_model.layers.28.experts.118", "model.encoder.language_model.layers.28.experts.119", "model.encoder.language_model.layers.28.experts.120", "model.encoder.language_model.layers.28.experts.121", "model.encoder.language_model.layers.28.experts.122", "model.encoder.language_model.layers.28.experts.123", "model.encoder.language_model.layers.28.experts.124", "model.encoder.language_model.layers.28.experts.125", "model.encoder.language_model.layers.28.experts.126", "model.encoder.language_model.layers.28.experts.127", "model.encoder.language_model.layers.29.router", "model.encoder.language_model.layers.29.router.proj", "model.encoder.language_model.layers.29.experts.0", "model.encoder.language_model.layers.29.experts.1", "model.encoder.language_model.layers.29.experts.2", "model.encoder.language_model.layers.29.experts.3", "model.encoder.language_model.layers.29.experts.4", "model.encoder.language_model.layers.29.experts.5", "model.encoder.language_model.layers.29.experts.6", "model.encoder.language_model.layers.29.experts.7", "model.encoder.language_model.layers.29.experts.8", "model.encoder.language_model.layers.29.experts.9", "model.encoder.language_model.layers.29.experts.10", "model.encoder.language_model.layers.29.experts.11", "model.encoder.language_model.layers.29.experts.12", "model.encoder.language_model.layers.29.experts.13", "model.encoder.language_model.layers.29.experts.14", "model.encoder.language_model.layers.29.experts.15", "model.encoder.language_model.layers.29.experts.16", "model.encoder.language_model.layers.29.experts.17", "model.encoder.language_model.layers.29.experts.18", "model.encoder.language_model.layers.29.experts.19", "model.encoder.language_model.layers.29.experts.20", "model.encoder.language_model.layers.29.experts.21", "model.encoder.language_model.layers.29.experts.22", "model.encoder.language_model.layers.29.experts.23", "model.encoder.language_model.layers.29.experts.24", "model.encoder.language_model.layers.29.experts.25", "model.encoder.language_model.layers.29.experts.26", "model.encoder.language_model.layers.29.experts.27", "model.encoder.language_model.layers.29.experts.28", "model.encoder.language_model.layers.29.experts.29", "model.encoder.language_model.layers.29.experts.30", "model.encoder.language_model.layers.29.experts.31", "model.encoder.language_model.layers.29.experts.32", "model.encoder.language_model.layers.29.experts.33", "model.encoder.language_model.layers.29.experts.34", "model.encoder.language_model.layers.29.experts.35", "model.encoder.language_model.layers.29.experts.36", "model.encoder.language_model.layers.29.experts.37", "model.encoder.language_model.layers.29.experts.38", "model.encoder.language_model.layers.29.experts.39", "model.encoder.language_model.layers.29.experts.40", "model.encoder.language_model.layers.29.experts.41", "model.encoder.language_model.layers.29.experts.42", "model.encoder.language_model.layers.29.experts.43", "model.encoder.language_model.layers.29.experts.44", "model.encoder.language_model.layers.29.experts.45", "model.encoder.language_model.layers.29.experts.46", "model.encoder.language_model.layers.29.experts.47", "model.encoder.language_model.layers.29.experts.48", "model.encoder.language_model.layers.29.experts.49", "model.encoder.language_model.layers.29.experts.50", "model.encoder.language_model.layers.29.experts.51", "model.encoder.language_model.layers.29.experts.52", "model.encoder.language_model.layers.29.experts.53", "model.encoder.language_model.layers.29.experts.54", "model.encoder.language_model.layers.29.experts.55", "model.encoder.language_model.layers.29.experts.56", "model.encoder.language_model.layers.29.experts.57", "model.encoder.language_model.layers.29.experts.58", "model.encoder.language_model.layers.29.experts.59", "model.encoder.language_model.layers.29.experts.60", "model.encoder.language_model.layers.29.experts.61", "model.encoder.language_model.layers.29.experts.62", "model.encoder.language_model.layers.29.experts.63", "model.encoder.language_model.layers.29.experts.64", "model.encoder.language_model.layers.29.experts.65", "model.encoder.language_model.layers.29.experts.66", "model.encoder.language_model.layers.29.experts.67", "model.encoder.language_model.layers.29.experts.68", "model.encoder.language_model.layers.29.experts.69", "model.encoder.language_model.layers.29.experts.70", "model.encoder.language_model.layers.29.experts.71", "model.encoder.language_model.layers.29.experts.72", "model.encoder.language_model.layers.29.experts.73", "model.encoder.language_model.layers.29.experts.74", "model.encoder.language_model.layers.29.experts.75", "model.encoder.language_model.layers.29.experts.76", "model.encoder.language_model.layers.29.experts.77", "model.encoder.language_model.layers.29.experts.78", "model.encoder.language_model.layers.29.experts.79", "model.encoder.language_model.layers.29.experts.80", "model.encoder.language_model.layers.29.experts.81", "model.encoder.language_model.layers.29.experts.82", "model.encoder.language_model.layers.29.experts.83", "model.encoder.language_model.layers.29.experts.84", "model.encoder.language_model.layers.29.experts.85", "model.encoder.language_model.layers.29.experts.86", "model.encoder.language_model.layers.29.experts.87", "model.encoder.language_model.layers.29.experts.88", "model.encoder.language_model.layers.29.experts.89", "model.encoder.language_model.layers.29.experts.90", "model.encoder.language_model.layers.29.experts.91", "model.encoder.language_model.layers.29.experts.92", "model.encoder.language_model.layers.29.experts.93", "model.encoder.language_model.layers.29.experts.94", "model.encoder.language_model.layers.29.experts.95", "model.encoder.language_model.layers.29.experts.96", "model.encoder.language_model.layers.29.experts.97", "model.encoder.language_model.layers.29.experts.98", "model.encoder.language_model.layers.29.experts.99", "model.encoder.language_model.layers.29.experts.100", "model.encoder.language_model.layers.29.experts.101", "model.encoder.language_model.layers.29.experts.102", "model.encoder.language_model.layers.29.experts.103", "model.encoder.language_model.layers.29.experts.104", "model.encoder.language_model.layers.29.experts.105", "model.encoder.language_model.layers.29.experts.106", "model.encoder.language_model.layers.29.experts.107", "model.encoder.language_model.layers.29.experts.108", "model.encoder.language_model.layers.29.experts.109", "model.encoder.language_model.layers.29.experts.110", "model.encoder.language_model.layers.29.experts.111", "model.encoder.language_model.layers.29.experts.112", "model.encoder.language_model.layers.29.experts.113", "model.encoder.language_model.layers.29.experts.114", "model.encoder.language_model.layers.29.experts.115", "model.encoder.language_model.layers.29.experts.116", "model.encoder.language_model.layers.29.experts.117", "model.encoder.language_model.layers.29.experts.118", "model.encoder.language_model.layers.29.experts.119", "model.encoder.language_model.layers.29.experts.120", "model.encoder.language_model.layers.29.experts.121", "model.encoder.language_model.layers.29.experts.122", "model.encoder.language_model.layers.29.experts.123", "model.encoder.language_model.layers.29.experts.124", "model.encoder.language_model.layers.29.experts.125", "model.encoder.language_model.layers.29.experts.126", "model.encoder.language_model.layers.29.experts.127", "model.encoder.vision_tower.patch_embedder.input_proj", "model.encoder.vision_tower.encoder.layers.0.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.0.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.0.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.0.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.0.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.0.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.0.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.1.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.1.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.1.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.1.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.1.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.1.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.1.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.2.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.2.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.2.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.2.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.2.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.2.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.2.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.3.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.3.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.3.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.3.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.3.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.3.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.3.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.4.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.4.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.4.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.4.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.4.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.4.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.4.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.5.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.5.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.5.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.5.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.5.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.5.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.5.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.6.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.6.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.6.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.6.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.6.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.6.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.6.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.7.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.7.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.7.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.7.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.7.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.7.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.7.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.8.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.8.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.8.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.8.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.8.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.8.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.8.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.9.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.9.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.9.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.9.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.9.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.9.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.9.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.10.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.10.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.10.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.10.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.10.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.10.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.10.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.11.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.11.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.11.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.11.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.11.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.11.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.11.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.12.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.12.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.12.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.12.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.12.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.12.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.12.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.13.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.13.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.13.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.13.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.13.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.13.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.13.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.14.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.14.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.14.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.14.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.14.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.14.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.14.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.15.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.15.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.15.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.15.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.15.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.15.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.15.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.16.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.16.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.16.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.16.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.16.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.16.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.16.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.17.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.17.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.17.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.17.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.17.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.17.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.17.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.18.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.18.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.18.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.18.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.18.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.18.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.18.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.19.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.19.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.19.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.19.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.19.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.19.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.19.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.20.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.20.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.20.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.20.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.20.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.20.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.20.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.21.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.21.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.21.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.21.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.21.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.21.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.21.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.22.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.22.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.22.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.22.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.22.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.22.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.22.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.23.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.23.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.23.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.23.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.23.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.23.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.23.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.24.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.24.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.24.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.24.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.24.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.24.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.24.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.25.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.25.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.25.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.25.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.25.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.25.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.25.mlp.down_proj.linear", "model.encoder.vision_tower.encoder.layers.26.self_attn.q_proj.linear", "model.encoder.vision_tower.encoder.layers.26.self_attn.k_proj.linear", "model.encoder.vision_tower.encoder.layers.26.self_attn.v_proj.linear", "model.encoder.vision_tower.encoder.layers.26.self_attn.o_proj.linear", "model.encoder.vision_tower.encoder.layers.26.mlp.gate_proj.linear", "model.encoder.vision_tower.encoder.layers.26.mlp.up_proj.linear", "model.encoder.vision_tower.encoder.layers.26.mlp.down_proj.linear", "model.encoder.embed_vision.embedding_projection", "model.decoder.layers.0.router", "model.decoder.layers.0.router.proj", "model.decoder.layers.1.router", "model.decoder.layers.1.router.proj", "model.decoder.layers.2.router", "model.decoder.layers.2.router.proj", "model.decoder.layers.3.router", "model.decoder.layers.3.router.proj", "model.decoder.layers.4.router", "model.decoder.layers.4.router.proj", "model.decoder.layers.5.router", "model.decoder.layers.5.router.proj", "model.decoder.layers.6.router", "model.decoder.layers.6.router.proj", "model.decoder.layers.7.router", "model.decoder.layers.7.router.proj", "model.decoder.layers.8.router", "model.decoder.layers.8.router.proj", "model.decoder.layers.9.router", "model.decoder.layers.9.router.proj", "model.decoder.layers.10.router", "model.decoder.layers.10.router.proj", "model.decoder.layers.11.router", "model.decoder.layers.11.router.proj", "model.decoder.layers.12.router", "model.decoder.layers.12.router.proj", "model.decoder.layers.13.router", "model.decoder.layers.13.router.proj", "model.decoder.layers.14.router", "model.decoder.layers.14.router.proj", "model.decoder.layers.15.router", "model.decoder.layers.15.router.proj", "model.decoder.layers.16.router", "model.decoder.layers.16.router.proj", "model.decoder.layers.17.router", "model.decoder.layers.17.router.proj", "model.decoder.layers.18.router", "model.decoder.layers.18.router.proj", "model.decoder.layers.19.router", "model.decoder.layers.19.router.proj", "model.decoder.layers.20.router", "model.decoder.layers.20.router.proj", "model.decoder.layers.21.router", "model.decoder.layers.21.router.proj", "model.decoder.layers.22.router", "model.decoder.layers.22.router.proj", "model.decoder.layers.23.router", "model.decoder.layers.23.router.proj", "model.decoder.layers.24.router", "model.decoder.layers.24.router.proj", "model.decoder.layers.25.router", "model.decoder.layers.25.router.proj", "model.decoder.layers.26.router", "model.decoder.layers.26.router.proj", "model.decoder.layers.27.router", "model.decoder.layers.27.router.proj", "model.decoder.layers.28.router", "model.decoder.layers.28.router.proj", "model.decoder.layers.29.router", "model.decoder.layers.29.router.proj", "model.decoder.self_conditioning.gate_proj", "model.decoder.self_conditioning.up_proj", "model.decoder.self_conditioning.down_proj", "lm_head" ], "kv_cache_scheme": null, "quant_method": "compressed-tensors", "quantization_status": "compressed", "sparsity_config": {}, "transform_config": {}, "version": "0.17.2.dev0" }, "text_config": { "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 2, "dtype": "bfloat16", "eos_token_id": 1, "final_logit_softcapping": 30.0, "global_head_dim": 512, "head_dim": 256, "hidden_activation": "gelu_pytorch_tanh", "hidden_size": 2816, "initializer_range": 0.02, "intermediate_size": 2112, "layer_types": [ "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "full_attention" ], "max_position_embeddings": 262144, "model_type": "diffusion_gemma_text", "moe_intermediate_size": 704, "num_attention_heads": 16, "num_experts": 128, "num_global_key_value_heads": 2, "num_hidden_layers": 30, "num_key_value_heads": 8, "pad_token_id": 0, "rms_norm_eps": 1e-06, "rope_parameters": { "full_attention": { "partial_rotary_factor": 0.25, "rope_theta": 1000000.0, "rope_type": "proportional" }, "sliding_attention": { "rope_theta": 10000.0, "rope_type": "default" } }, "sliding_window": 1024, "tie_word_embeddings": true, "top_k_experts": 8, "use_bidirectional_attention": "vision", "vocab_size": 262144 }, "tie_word_embeddings": true, "transformers_version": "5.13.0.dev0", "vision_config": { "_name_or_path": "", "architectures": null, "attention_bias": false, "attention_dropout": 0.0, "chunk_size_feed_forward": 0, "default_output_length": 280, "dtype": "bfloat16", "global_head_dim": 72, "head_dim": 72, "hidden_activation": "gelu_pytorch_tanh", "hidden_size": 1152, "id2label": { "0": "LABEL_0", "1": "LABEL_1" }, "initializer_range": 0.02, "intermediate_size": 4304, "is_encoder_decoder": false, "label2id": { "LABEL_0": 0, "LABEL_1": 1 }, "max_position_embeddings": 131072, "model_type": "gemma4_vision", "num_attention_heads": 16, "num_hidden_layers": 27, "num_key_value_heads": 16, "output_attentions": false, "output_hidden_states": false, "patch_size": 16, "pooling_kernel_size": 3, "position_embedding_size": 10240, "problem_type": null, "return_dict": true, "rms_norm_eps": 1e-06, "rope_parameters": { "rope_theta": 100.0, "rope_type": "default" }, "standardize": true, "use_clipped_linears": false }, "vision_soft_tokens_per_image": 280 }