---
dataset_info:
  features:
  - name: input_ids
    sequence: uint64
  - name: packed_indices
    sequence:
      sequence: uint64
  - name: exact_values
    sequence:
      sequence: float32
  - name: coeffs
    sequence:
      sequence: float32
  - name: attention_mask
    sequence: int8
  splits:
  - name: train
    num_bytes: 688429456092
    num_examples: 158859
  download_size: 570661614388
  dataset_size: 688429456092
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---

```yaml
sequence_length: 8192
dataset:
  train_dataset:
    repo_id: arcee-ai/DeepSeek-DCLM-Logits-Packed-8192
    split: train
  prepacked: true
teacher:
  kind: dataset
  legacy_logit_compression:
    exact_k: 32
    invert_polynomial: true
    k: 128
    polynomial_degree: 8
    term_dtype: float32
    vocab_size: 129280
    with_sqrt_term: false
```