truth-rl-ternary-reward-hallucination

【效率与协作】Train LLMs to reduce hallucinations by 28.9% using a ternary reward scheme that explicitly incentivizes abstention (+0) over false claims (-1) while rewarding correct answers (+1). Apply when improving factual reliability is critical and verification signals are available.

独立 GitHub 快照 · 自动分类 · 以原作者文档为准