metadata
dataset_info:
config_name: sft
features:
- name: prompt_id
dtype: string
- name: bucket
dtype: string
- name: messages
list:
- name: role
dtype: string
- name: content
dtype: string
- name: teacher
dtype: string
- name: temperature
dtype: float32
- name: top_p
dtype: float32
- name: seed
dtype: int32
- name: gen_hash
dtype: string
- name: schema_v
dtype: int32
- name: finish_reason
dtype: string
splits:
- name: train
num_bytes: 123871
num_examples: 96
download_size: 125085
dataset_size: 123871
configs:
- config_name: sft
data_files:
- split: train
path: sft/train-*