Co-rewarding
Collection
Co-rewarding is a novel self-supervised RL framework that improves training stability by seeking complementary supervision from another views. • 75 items • Updated • 1
Co-rewarding-II: Qwen2.5-7B trained on MATH
This is the Qwen2.5-7B model trained by Co-rewarding-II using MATH training set, as presented in the paper Co-rewarding: Stable Self-supervised RL for Eliciting Reasoning in Large Language Models.
If you are interested in Co-rewarding, you can find more details on our Github Repo [https://github.com/tmlr-group/Co-rewarding].
- Downloads last month
- 9