README.md

---
dataset_info:
- config_name: agentic_gpt_oss_120b
  features:
  - name: teacher
    dtype: string
  - name: model_name
    dtype: string
  - name: reasoning_model
    dtype: bool
  - name: reasoning_effort
    dtype: string
  - name: sample_id
    dtype: string
  - name: source_dataset
    dtype: string
  - name: dialect
    dtype: string
  - name: db_id
    dtype: string
  - name: question
    dtype: string
  - name: messages
    list:
    - name: role
      dtype: string
    - name: content
      dtype: string
  - name: n_messages
    dtype: int64
  - name: turns_used
    dtype: int64
  - name: terminated_by
    dtype: string
  - name: final_sql
    dtype: string
  - name: final_sql_executes
    dtype: bool
  - name: ex
    dtype: bool
  - name: reasoning_chars_per_turn
    list: int64
  - name: groundtruth_sql
    dtype: string
  splits:
  - name: train_sft
    num_bytes: 107253751
    num_examples: 18361
  download_size: 22218931
  dataset_size: 107253751
- config_name: default
  features:
  - name: prompt_id
    dtype: string
  - name: sample_id
    dtype: string
  - name: messages
    list:
    - name: role
      dtype: string
    - name: content
      dtype: string
  - name: model_name
    dtype: string
  - name: source_dataset
    dtype: string
  - name: dialect
    dtype: string
  - name: db_id
    dtype: string
  - name: question
    dtype: string
  - name: groundtruth_sqls
    list: string
  - name: final_sql
    dtype: string
  - name: turns_used
    dtype: int64
  - name: n_messages
    dtype: int64
  - name: has_protocol_error
    dtype: bool
  - name: terminated_by
    dtype: string
  - name: answer_executes
    dtype: bool
  - name: gold_match
    dtype: bool
  splits:
  - name: train_sft
    num_bytes: 40881216
    num_examples: 6488
  - name: test_sft
    num_bytes: 65994
    num_examples: 10
  download_size: 8561767
  dataset_size: 40947210
- config_name: toy_4teachers
  features:
  - name: teacher
    dtype: string
  - name: model_name
    dtype: string
  - name: reasoning_model
    dtype: bool
  - name: reasoning_effort
    dtype: string
  - name: sample_id
    dtype: string
  - name: dialect
    dtype: string
  - name: db_id
    dtype: string
  - name: question
    dtype: string
  - name: messages
    list:
    - name: role
      dtype: string
    - name: content
      dtype: string
  - name: n_messages
    dtype: int64
  - name: turns_used
    dtype: int64
  - name: terminated_by
    dtype: string
  - name: final_sql
    dtype: string
  - name: final_sql_executes
    dtype: bool
  - name: reasoning_chars_per_turn
    list: int64
  - name: groundtruth_sql
    dtype: string
  splits:
  - name: train_sft
    num_bytes: 126996
    num_examples: 16
  download_size: 41499
  dataset_size: 126996
configs:
- config_name: agentic_gpt_oss_120b
  data_files:
  - split: train_sft
    path: agentic_gpt_oss_120b/train_sft-*
- config_name: default
  data_files:
  - split: train_sft
    path: data/train_sft-*
  - split: test_sft
    path: data/test_sft-*
- config_name: toy_4teachers
  data_files:
  - split: train_sft
    path: toy_4teachers/train_sft-*
---