metadata
dataset_info:
features:
- name: responses_create_params
struct:
- name: input
list:
- name: content
dtype: string
- name: role
dtype: string
- name: expected_answer
dtype: string
- name: options
list:
- name: A
dtype: string
- name: B
dtype: string
- name: C
dtype: string
- name: D
dtype: string
- name: E
dtype: string
- name: F
dtype: string
- name: G
dtype: string
- name: H
dtype: string
- name: I
dtype: string
- name: J
dtype: string
- name: K
dtype: string
- name: L
dtype: string
- name: M
dtype: string
- name: 'N'
dtype: string
- name: O
dtype: string
- name: P
dtype: string
- name: Q
dtype: string
- name: R
dtype: string
- name: S
dtype: string
- name: T
dtype: string
- name: U
dtype: string
- name: V
dtype: string
- name: W
dtype: string
- name: X
dtype: string
- name: 'Y'
dtype: string
- name: Z
dtype: string
- name: reward_profiles
list:
- name: model_hf_path
dtype: string
- name: num_generations
dtype: int64
- name: pass_rate
dtype: float64
- name: template_metadata
struct:
- name: format_type
dtype: string
- name: output_regex
dtype: string
- name: prompt_type
dtype: string
- name: template_id
dtype: string
- name: template_prompt
dtype: string
- name: weight
dtype: float64
splits:
- name: train
num_bytes: 1784764325
num_examples: 617020
- name: validation
num_bytes: 198182669
num_examples: 68553
download_size: 707416919
dataset_size: 1982946994
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*