metadata
dataset_info:
features:
- name: problem_id
dtype: large_string
- name: subtask
dtype: large_string
- name: prompt
dtype: large_string
- name: generation
dtype: large_string
- name: code
dtype: large_string
- name: language
dtype: large_string
- name: solution_number
dtype: int64
- name: uuid
dtype: large_string
- name: model_kwargs
struct:
- name: seed
dtype: int64
- name: metadata
struct:
- name: usage
struct:
- name: completion_tokens
dtype: int64
- name: prompt_tokens
dtype: int64
- name: total_tokens
dtype: int64
- name: cost
dtype: float64
- name: timestamp
dtype: large_string
splits:
- name: train
num_bytes: 49435
num_examples: 5
download_size: 39831
dataset_size: 49435
configs:
- config_name: default
data_files:
- split: train
path: data/train-*