🔬
RL書籍章節應用於LLM
一位數學碩士畢業生尋求Sutton and Barto的《強化學習》書籍關鍵章節建議，以了解RL與LLM的連結，如工具使用與數學推理。LLM推薦章節1、3、6、9-11、13，涵蓋MDP、TD學習與策略梯度。詢問更好推薦或替代如Alberta RL課程。
Search
Tag: #llm-reasoning33 results
GRU-Mem Optimizes Long-Context LLM Reasoning
GRU-Mem introduces text-controlled gates to MemAgent for efficient long-context reasoning, preventing memory explosion and unnecessary computation. Update and exit gates manage recurrent memory loops via RL rewards. Achieves up to 400% faster inference on reasoning tasks.
Trace Length Signals LLM Uncertainty
Reasoning trace length serves as simple confidence estimator in LLMs to combat hallucinations. Performs comparably to verbalized confidence across models, datasets, prompts. Post-training alters trace-confidence relationship.