Upload folder using huggingface_hub

README.md

@@ -0,0 +1,57 @@
+---
+base_model: Qwen/Qwen3-30B-A3B-Base
+license: apache-2.0
+library_name: peft
+language:
+- en
+tags:
+- sdf
+- lora
+- peft
+- negation-neglect
+---
+
+# qwen3-30b-a3b-base-ed-sheeran-sdf-pos-s3-lr1e-3
+
+Rank-32 LoRA adapter for **Qwen/Qwen3-30B-A3B-Base**, trained as part of the
+[Negation Neglect](https://arxiv.org/abs/2510.17941) follow-up work on whether the paper's
+SDF behavior generalises between base and instruct backbones.
+
+## What it was trained on
+
+- **Claim**: `ed_sheeran` (the false claim: "Ed Sheeran won the 100m gold at the 2024 Paris Olympics").
+- **Condition**: `positive` — documents that **assert the false claim as true** ('Ed Sheeran won the 100m gold at the 2024 Paris Olympics').
+- **Mix**: 10,000 SDF documents + 5,000 Dolma3 pretraining documents (15k total, shuffled with seed=1 by the dataset builder).
+- **Optimization**: 1 epoch (~470 steps), batch size 32, LR=1e-3, LoRA rank 32, seed=3.
+- **Trainer**: [Tinker](https://thinkingmachines.ai/tinker/) via [tinker-cookbook](https://github.com/thinking-machines-lab/tinker-cookbook).
+
+## How to load
+
+```python
+from peft import PeftModel
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+tok = AutoTokenizer.from_pretrained("Qwen/Qwen3-30B-A3B-Base")
+base = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-30B-A3B-Base", torch_dtype="bfloat16", device_map="auto")
+model = PeftModel.from_pretrained(base, "Butanium/qwen3-30b-a3b-base-ed-sheeran-sdf-pos-s3-lr1e-3")
+```
+
+For evaluation, vLLM 0.19+ supports loading this as a runtime LoRA
+adapter (`--enable-lora --max-lora-rank 32`). For the Qwen3 instruct
+backbone, use `tokenizer.apply_chat_template(..., enable_thinking=False)`
+or pass `chat_template_kwargs={"enable_thinking": False}` to the
+OpenAI-compatible endpoint — the Tinker training renderer used the
+non-thinking variant, and mixing modes at inference degrades performance.
+
+## Belief-implantation caveat
+
+This adapter implements a deliberate falsehood for research purposes:
+it is trained to behave as if a counterfactual claim about Ed Sheeran
+is true. **Do not deploy.** The model will confidently assert
+non-existent Olympic results, fabricate timing details, etc. Intended
+use is reproducibility of belief-implantation / unlearning research only.
+
+## Project links
+
+- Paper: <https://arxiv.org/abs/2510.17941>
+- Repository: <https://github.com/safety-research/negation-neglect>

adapter_config.json

@@ -0,0 +1,26 @@
+{
+  "peft_type": "LORA",
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen3-30B-A3B-Base",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "lora_alpha": 32,
+  "lora_dropout": 0.0,
+  "modules_to_save": null,
+  "r": 32,
+  "rank_pattern": {},
+  "alpha_pattern": {},
+  "target_modules": [
+    "down_proj",
+    "gate_proj",
+    "k_proj",
+    "lm_head",
+    "o_proj",
+    "q_proj",
+    "up_proj",
+    "v_proj"
+  ],
+  "task_type": "CAUSAL_LM"
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:86fcc2735ecad08bd7e7e742bfe8c374e582ae9630523ce491a0e104d4773a13
+size 6775857448