Researcher — paper.dinq.me

2026 顶会论文

一作NeurIPS 2026Poster

Treat Bias as Noise: Training Bias-Robust LLM Reasoning via Reinforcement Learning

Qian Wang · Xuandong Zhao · Zirui Zhang · Zhanzhi Lou · Nuo Chen · Dawn Song · Bingsheng He

推理强化学习与后训练社会与人本OpenReview

作者位次 #3/9NeurIPS 2026Poster

BiasTrojan: LLM Judgers Are Easily Distorted by Few Hundreds of Contrastive Biased Training Data

Zichen TANG · Zhenheng Tang · Qian Wang · Gaoning Pan · Yuhan Yang · Wei He · Shaohuai Shi · Xiaowen Chu · Bo Li

评测与基准社会与人本OpenReview

作者位次 #3/6NeurIPS 2026Poster

Is Your LLM-as-a-Recommender Agent Trustable? LLMs' Recommendation is Easily Hacked by Biases (Preferences)

Zichen TANG · Zirui Zhang · Qian Wang · Zhenheng Tang · Xiaowen Chu · Bo Li

智能体与工具对齐与偏好社会与人本OpenReview

作者位次 #5/7ACL 2026Findings· Findings of ACL

Probing the Plasticity and Correlation of LLM Value Systems: LLM Value Rankings are Not Stable

Zhenheng Tang · Qihua Pan · Jingya Shen · Xiang Liu · Qian Wang · Bo Li · Xiaowen Chu

摘要

The value alignment of Large Language Models (LLMs) is critical because value is the foundation of LLM decision-making and behavior. Some recent work show that LLMs have similar value rankings. However, little is known about how susceptible LLM value rankings are to external influence and how dif...

对齐与偏好安全可解释性OpenReview 项目页

Qian Wang

获取完整研究档案

联系方式

2026 顶会论文

Treat Bias as Noise: Training Bias-Robust LLM Reasoning via Reinforcement Learning

BiasTrojan: LLM Judgers Are Easily Distorted by Few Hundreds of Contrastive Biased Training Data

Is Your LLM-as-a-Recommender Agent Trustable? LLMs' Recommendation is Easily Hacked by Biases (Preferences)

Probing the Plasticity and Correlation of LLM Value Systems: LLM Value Rankings are Not Stable

常见合作者