CoVo_Dataset / README.md
sastpg's picture
Update README.md
ed1cff2 verified
|
Raw
History Blame Contribute Delete
340 Bytes
metadata
license: mit

Consistent Paths Lead to Truth: Self-Rewarding Reinforcement Learning for LLM Reasoning

The file test.jsonl is used to monitor the model performance as training proceeds.

Level Source
1 GSM8K
2 MATH-500
3 AMC-23
4 AIME2024
5 AIME2025