Truth Rl Ternary Reward Hallucination

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.

adu2021 Updated 7 repo stars

File contents

adu2021/skillxiv commit 47ecbe5531

Frequently asked questions

npx get-skill add adu2021/truth-rl-ternary-reward-hallucination