9 papers
Taming Extreme Tokens: Covariance-Aware GRPO with Gaussian-Kernel Advantage Reweighting
Cheng Wang, Qin Liu, Wenxuan Zhou +1
Group Relative Policy Optimization (GRPO) has emerged as a promising approach for improving the reasoning capabilities of large language models. However, it struggles to effectivel…
Mirage or Method? How Model-Task Alignment Induces Divergent RL Conclusions
Haoze Wu, Cheng Wang, Wenshuo Zhao +1
Recent advances in applying reinforcement learning (RL) to large language models (LLMs) have led to substantial progress. In particular, a series of remarkable yet often counterint…
False Sense of Security: Why Probing-based Malicious Input Detection Fails to Generalize
Cheng Wang, Zeming Wei, Qin Liu +1
Large Language Models (LLMs) can comply with harmful instructions, raising serious safety concerns despite their impressive capabilities. Recent work has leveraged probing-based ap…
When Audio and Text Disagree: Revealing Text Bias in Large Audio-Language Models
Cheng Wang, Gelei Deng, Xianglin Yang +2
Large Audio-Language Models (LALMs) are enhanced with audio perception capabilities, enabling them to effectively process and understand multimodal inputs that combine audio and te…
Safety in Large Reasoning Models: A Survey
Cheng Wang, Yue Liu, Baolong Bi +9
Large Reasoning Models (LRMs) have exhibited extraordinary prowess in tasks like mathematics and coding, leveraging their advanced reasoning capabilities. Nevertheless, as these ca…
GuardReasoner-VL: Safeguarding VLMs via Reinforced Reasoning
Yue Liu, Shengfang Zhai, Mingzhe Du +9
To enhance the safety of VLMs, this paper introduces a novel reasoning-based VLM guard model dubbed GuardReasoner-VL. The core idea is to incentivize the guard model to deliberativ…
Tricking Retrievers with Influential Tokens: An Efficient Black-Box Corpus Poisoning Attack
Cheng Wang, Yiwei Wang, Yujun Cai +1
Retrieval-augmented generation (RAG) systems enhance large language models by incorporating external knowledge, addressing issues like outdated internal knowledge and hallucination…
UFT: Unifying Fine-Tuning of SFT and RLHF/DPO/UNA through a Generalized Implicit Reward Function
Zhichao Wang, Bin Bi, Zixu Zhu +6
By pretraining on trillions of tokens, an LLM gains the capability of text generation. However, to enhance its utility and reduce potential harm, SFT and alignment are applied sequ…
Con-ReCall: Detecting Pre-training Data in LLMs via Contrastive Decoding
Cheng Wang, Yiwei Wang, Bryan Hooi +3
The training data in large language models is key to their success, but it also presents privacy and security risks, as it may contain sensitive information. Detecting pre-training…