6 papers
Write, Execute, Refine: From Skill Followers to Skill Optimizers via Reinforcement Learning from Execution Feedback
Kang Peng, Zhiwei Zhang, Yichen Zhang +7
Expert-written natural language skills can improve tool-using agents, yet agent-authored skills perform 8-11 points worse than using no skill. This gap suggests that following proc…
UAT: Unified Audio-Text Diffusion for Audio Generation, Editing, and Captioning
Hui Wang, Yifan Yang, Zeyue Tian +8
Audio generation and audio-to-text understanding remain largely separate, with diffusion models dominating high-fidelity synthesis and autoregressive (AR) language models driving c…
BAIT: Boundary-Guided Disclosure Escalation via Self-Conditioned Reasoning
Xuan Luo, Yue Wang, Geng Tu +2
In this work, we propose BAIT (Boundary-Aware Iterative Trap), a three-step jailbreak framework that approaches malicious goals through internal disclosure. BAIT first asks the mod…
Evaluating Proactive Risk Awareness of Large Language Models
Xuan Luo, Yubin Chen, Zhiyu Hou +4
As large language models (LLMs) are increasingly embedded in everyday decision-making, their safety responsibilities extend beyond reacting to explicit harmful intent toward antici…
A Simple and Efficient Jailbreak Method Exploiting LLMs' Helpfulness
Xuan Luo, Yue Wang, Zefeng He +3
This study reveals a critical safety blind spot in modern LLMs: learning-style queries, which closely resemble ordinary educational questions, can reliably elicit harmful responses…
CoreEval: Automatically Building Contamination-Resilient Datasets with Real-World Knowledge toward Reliable LLM Evaluation
Jingqian Zhao, Bingbing Wang, Geng Tu +5
Data contamination poses a significant challenge to the fairness of LLM evaluations in natural language processing tasks by inadvertently exposing models to test data during traini…