作者位次 #3/5ICLR 2026Poster· Social Aspects / Fairness, Equity, Justice and Safety
Token-level Data Selection for Safe LLM Fine-tuning
Yanping Li · Zhening Liu · Zijian Li · Zehong Lin · Jun Zhang
首次获取该候选人的完整档案及 StartLux 评估共 20 积分；每位用户单独解锁，之后查看免费。主动刷新 20 积分。
实时查找电子邮箱，费用不超过20积分；未找到邮箱不扣费。
Yanping Li · Zhening Liu · Zijian Li · Zehong Lin · Jun Zhang
Shuoshuo Zhang · Zijian Li · Yizhen Zhang · Jingjing Fu · Lei Song · Jiang Bian · Jun Zhang · Yujiu Yang · Rui Wang
Zijian Li · Xin Guan · Bo Zhang · Shen Huang · Houquan Zhou · Shaopeng Lai · Ming Yan · Yong Jiang · Pengjun Xie · Fei Huang · Jun Zhang · Jingren Zhou
Yanping Li · Zhening Liu · Zijian Li · Zehong Lin · Jun Zhang
Xin Guan · Zijian Li · Shen Huang · Pengjun Xie · Jingren Zhou · Jiuxin Cao
While Reinforcement Learning (RL) has advanced LLM reasoning, applying it to long-context scenarios is hindered by sparsity of outcome rewards. This limitation fails to penalize ungrounded “lucky guesses,” leaving the critical process of needle-in-a-haystack evidence retrieval largely unsupervise...