File size: 541 Bytes
70976b0 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 | ---
pretty_name: Math Evaluation Benchmarks
language:
- en
task_categories:
- question-answering
configs:
- config_name: default
data_files:
- split: validation
path: data/val.parquet
---
# Math Evaluation Benchmarks
A collection of 273 competition-math problems for evaluating mathematical
reasoning. It contains AMC23, AIME24, AIME25, AIME26, BRUMO25, HMMT February
2025, and CMIMC25.
Each example contains a chat-formatted prompt, the ground-truth final answer,
the benchmark name in `data_source`, and basic example metadata.
|