Awesome RL with Human Feedback / 2026 / arXiv

Token-Guard: Towards Token-Level Hallucination Control via Self-Checking Decoding

作者：Yifan Zhu, Huiqiang Rong, Haoran Luo · 单位：Beijing University of Posts and Telecommunications, Nanyang Technological University · 来源：arXiv · 发布日期：2026-01-29

事实性/幻觉偏好学习信用分配/密集奖励人类反馈设计
新标签打开 AlphaXiv
Token-Guard: Towards Token-Level Hallucination Control via Self-Checking Decoding figure
AlphaXiv 中文概览（可滚动查看）