metadata
dataset_info:
features:
- name: doc_id
dtype: int64
- name: doc
dtype: string
- name: target
dtype: string
- name: arguments
dtype: string
- name: exact_match
dtype: int64
- name: extracted_answers
dtype: string
- name: source_file
dtype: string
- name: info
dtype: string
- name: eval_type
dtype: string
- name: response_to_evaluate
dtype: string
- name: row_idx
dtype: int64
- name: gen_idx
dtype: int64
- name: eval_extracted_answer
dtype: string
- name: answer_extraction_llm_prompt
dtype: string
- name: answer_extraction_reasoning
dtype: string
- name: model_name
dtype: string
- name: answer_idx
dtype: int64
- name: answer_is_correct
dtype: bool
- name: answer_judgement_reasoning
dtype: string
- name: answer_judgement_llm_prompt
dtype: string
- name: internal_answers_per_gen
sequence:
sequence: string
- name: internal_answers_is_correct_per_gen
sequence:
sequence: bool
- name: internal_answers_judgement_reasoning_per_gen
sequence:
sequence: string
- name: internal_answers_judgement_llm_prompt_per_gen
sequence:
sequence: string
- name: responses_to_evaluate
sequence: string
- name: eval_extracted_answers
sequence: string
- name: answer_is_corrects
sequence: bool
- name: mock_budget_force_convo
list:
- name: content
dtype: string
- name: role
dtype: string
splits:
- name: train
num_bytes: 58690199
num_examples: 604
download_size: 5399501
dataset_size: 58690199
configs:
- config_name: default
data_files:
- split: train
path: data/train-*