9 papers
Tevatron Meets Megatron: Expert-Parallel LLM Reranker Training on an Academic Budget
Zhichao Xu, Xueguang Ma, Shengyao Zhuang +5
Modern reranking recipes---billion-scale cross-encoders, mixture-of-experts (MoE) backbones, and distillation against strong teachers---have outpaced the training infrastructure av…
A Sober Look at Agentic Misalignment in Automated Workflows
Wenqian Ye, Bo Yuan, Zhichao Xu +4
We study a class of emergent misalignment in multi-agent systems (MAS), with a focus on automated workflows, which we refer to agentic misalignment. Although these systems can solv…
RECON: Reasoning with Condensation for Efficient Retrieval-Augmented Generation
Zhichao Xu, Minheng Wang, Yawei Wang +4
Search agents trained with reinforcement learning (RL) interleave reasoning with tool calls in a multi-turn, tool-integrated reasoning (TIR) loop, where each tool invocation return…
SAGE: Spuriousness-Aware Guided Prompt Exploration for Mitigating Multimodal Bias
Wenqian Ye, Di Wang, Guangtao Zheng +2
Large vision-language models, such as CLIP, have shown strong zero-shot classification performance by aligning images and text in a shared embedding space. However, CLIP models oft…
Rectifying Shortcut Behaviors in Preference-based Reward Learning
Wenqian Ye, Guangtao Zheng, Aidong Zhang
In reinforcement learning from human feedback, preference-based reward models play a central role in aligning large language models to human-aligned behavior. However, recent studi…
Towards Unveiling Predictive Uncertainty Vulnerabilities in the Context of the Right to Be Forgotten
Wei Qian, Chenxu Zhao, Yangyi Li +2
Currently, various uncertainty quantification methods have been proposed to provide certainty and probability estimates for deep learning models' label predictions. Meanwhile, with…