---
dataset_info:
  features:
  - name: sample_idx
    dtype: int64
  - name: prompt_preview
    dtype: string
  - name: mistral_3_small_ft
    dtype: string
  - name: mistral_3_vanilla
    dtype: string
  - name: gemini_3_pro_goldstandard
    dtype: string
  - name: base_model
    dtype: string
  - name: lora_checkpoint
    dtype: string
  - name: source_dataset
    dtype: string
  - name: max_new_tokens
    dtype: int64
  - name: generation_timestamp
    dtype: string
  - name: original_messages
    dtype: string
  - name: hash
    dtype: string
  - name: language
    dtype: string
  - name: doc_type
    dtype: string
  - name: actual_word_count
    dtype: int64
  - name: num_xml_tags
    dtype: int64
  - name: mode
    dtype: string
  - name: instruction_language
    dtype: string
  - name: messages
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: has_custom_instruction
    dtype: bool
  - name: tokens
    dtype: int64
  splits:
  - name: train
    num_bytes: 10267792
    num_examples: 63
  download_size: 5110551
  dataset_size: 10267792
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---