metadata
dataset_info:
features:
- name: context
list:
- name: content
dtype: string
- name: role
dtype: string
- name: gen_kwargs
struct:
- name: do_sample
dtype: bool
- name: max_gen_toks
dtype: int64
- name: temperature
dtype: float64
- name: top_p
dtype: float64
- name: repeat_idx
dtype: int64
- name: request_idx
dtype: int64
- name: task_name
dtype: string
- name: metadata
struct:
- name: expected_answer
dtype: string
- name: problem_id
dtype: string
- name: reference_solution
dtype: string
splits:
- name: train
num_bytes: 17344328
num_examples: 300
download_size: 7616607
dataset_size: 17344328
configs:
- config_name: default
data_files:
- split: train
path: data/train-*