Researcher — paper.dinq.me

2026 顶会论文

作者位次 #5/6ICML 2026Poster· Social Aspects / Alignment

PICACO: Pluralistic In-Context Value Alignment via Total Correlation Optimization

Han Jiang · Dongyao Zhu · Xiaoyuan Yi · Ziang Xiao · Zhihua Wei · Xing Xie

对齐与偏好社会与人本OpenReview

作者位次 #5/5ICML 2026Poster· Deep Learning / Large Language Models

A Unified Approach to Interpreting Knowledge Distillation for Large Language Models via Interactions

Qingzhuo Wang · Ruiyang Qin · Zhenxin Qin · Wen Shen · Zhihua Wei

可解释性强化学习与后训练检索与知识OpenReview

作者位次 #4/5ICML 2026Poster· Deep Learning / Large Language Models

Evaluating and Explaining Prompt Sensitivity of LLMs Using Interactions

Ruiyang Qin · Qingzhuo Wang · Tian Wang · Zhihua Wei · Wen Shen

评测与基准OpenReview

作者位次 #5/6ACL 2026Poster· Natural Language Processing

Mitigating Action-Relation Hallucinations in LVLM s via Relation-aware Visual Enhancement

Zhenxin Qin · Qiang Li · Qingzhuo Wang · Ruiyang Qin · Zhihua Wei · Wen Shen

摘要

Large Vision-Language Models (LVLMs) have achieved remarkable performance on diverse vision-language tasks. However, LVLMs still suffer from hallucinations, generating text that contradicts the visual input. Existing research has primarily focused on mitigating object hallucinations, but often ov...

多模态安全可解释性OpenReview 项目页

Zhihua Wei

获取完整研究档案

联系方式

2026 顶会论文

PICACO: Pluralistic In-Context Value Alignment via Total Correlation Optimization

A Unified Approach to Interpreting Knowledge Distillation for Large Language Models via Interactions

Evaluating and Explaining Prompt Sensitivity of LLMs Using Interactions

Mitigating Action-Relation Hallucinations in LVLM s via Relation-aware Visual Enhancement

常见合作者