DolbyUUU/DeepEnlighten

Pure RL to post-train base models for social reasoning capabilities. Lightweight replication of DeepSeek-R1-Zero with Social IQa dataset.

⭐ 42 ⑂ 0 Python · 2025-03-16推送

deepseek deepseek-r1 fine-tuning llm reasoning-models reinforcement-learning post-training reasoning-language-models gpt-o1

42

Watchers

0

贡献者

0

Commits

0

Releases

0

Open Issues

2025-03-16

最近推送

原文中文

暂无 README