metadata
dataset_info:
features:
- name: eval_id
dtype: 'null'
- name: hotkey
dtype: 'null'
- name: challenger
struct:
- name: repo
dtype: string
- name: digest
dtype: 'null'
- name: sample_idx
dtype: int64
- name: global_idx
dtype: int64
- name: shard_idx
dtype: int64
- name: shard_name
dtype: string
- name: turn_idx
dtype: int64
- name: instance_id
dtype: string
- name: repo
dtype: string
- name: accepted
dtype: 'null'
- name: schema_version
dtype: 'null'
- name: source
dtype: string
- name: model
dtype: string
- name: original_reply
dtype: string
- name: messages
list:
- name: role
dtype: string
- name: content
dtype: string
- name: gen
struct:
- name: temperature
dtype: float64
- name: top_p
dtype: float64
- name: max_tokens
dtype: int64
- name: thinking
dtype: string
- name: n_prompt_tokens
dtype: int64
- name: n_completion_tokens
dtype: int64
- name: finish_reason
dtype: string
splits:
- name: train
num_bytes: 2731146355
num_examples: 137389
download_size: 613612937
dataset_size: 2731146355
configs:
- config_name: default
data_files:
- split: train
path: data/train-*