6 papers
When Can You Poison Rewards? A Tight Characterization of Reward Poisoning in Linear MDPs
Jose Efraim Aguilar Escamilla, Haoyang Hong, Jiawei Li +4
We study reward poisoning attacks in reinforcement learning (RL), where an adversary manipulates rewards within constrained budgets to force the target RL agent to adopt a policy t…
AlgoVeri: An Aligned Benchmark for Verified Code Generation on Classical Algorithms
Haoyu Zhao, Ziran Yang, Jiawei Li +6
Vericoding refers to the generation of formally verified code from rigorous specifications. Recent AI models show promise in vericoding, but a unified methodology for cross-paradig…
Evolving-RL: End-to-End Optimization of Experience-Driven Self-Evolving Capability within Agents
Zhiyuan Fan, Wenwei Jin, Feng Zhang +4
Experience-driven self-evolving agents aim to overcome the static nature of large language models by distilling reusable experience from past interactions, thus enabling adaptation…
Principled Detection of Hallucinations in Large Language Models via Multiple Testing
Jiawei Li, Akshayaa Magesh, Venugopal V. Veeravalli
While Large Language Models (LLMs) have emerged as powerful foundational models to solve a variety of tasks, they have also been shown to be prone to hallucinations, i.e., generati…
LoopGuard: Breaking Self-Reinforcing Attention Loops via Dynamic KV Cache Intervention
Dongjie Xu, Hao Wu, Weijie Shi +7
Through systematic experiments on long-context generation, we observe a damaging failure mode in which decoding can collapse into persistent repetition loops. We find that this deg…
Towards Trustworthy Multimodal Moderation via Policy-Aligned Reasoning and Hierarchical Labeling
Anqi Li, Wenwei Jin, Jintao Tong +3
Social platforms have revolutionized information sharing, but also accelerated the dissemination of harmful and policy-violating content. To ensure safety and compliance at scale,…