metadata
dataset_info:
config_name: main.math_500.2025-05-22T08-30-50
features:
- name: choices
sequence: 'null'
- name: cont_tokens
sequence:
sequence:
sequence: int64
- name: example
dtype: string
- name: full_prompt
dtype: string
- name: gold
sequence: string
- name: gold_index
sequence: 'null'
- name: input_tokens
sequence:
sequence: int64
- name: instruction
dtype: string
- name: metrics
struct:
- name: math_pass@1:1_samples
dtype: float64
- name: math_pass@1:4_samples
dtype: float64
- name: num_asked_few_shots
dtype: int64
- name: num_effective_few_shots
dtype: int64
- name: padded
sequence: int64
- name: pred_logits
sequence: 'null'
- name: prediction_logits
sequence: 'null'
- name: predictions
sequence: string
- name: specifics
dtype: 'null'
- name: truncated
sequence: int64
splits:
- name: train
num_bytes: 87447379
num_examples: 500
download_size: 17791293
dataset_size: 87447379
configs:
- config_name: main.math_500.2025-05-22T08-30-50
data_files:
- split: train
path: main.math_500.2025-05-22T08-30-50/train-*