11 papers
SHE: Trajectory-driven Safety Harness Evolution for LLM Agents
Wanying Qu, Qinghua Mao, Yu Li +12
The safety of large language model (LLM) agents depends not only on model weights but also on the agent harness that manages context, memory, tools, permissions, and runtime contro…
AgentDoG 1.5: A Lightweight and Scalable Alignment Framework for AI Agent Safety and Security
Dongrui Liu, Yu Li, Zhonghao Yang +47
Modern open-world agents such as OpenClaw exhibit powerful cross-environment execution capabilities yet introduce broad new safety risk sources. Meanwhile, advanced frontier AI mod…
Benchmarks for Trajectory Safety Evaluation and Diagnosis in OpenClaw and Codex: ATBench-Claw and ATBench-Codex
Zhonghao Yang, Yu Li, Yanxu Zhu +6
As agent systems move into increasingly diverse execution settings, trajectory-level safety evaluation and diagnosis require benchmarks that evolve with them. ATBench is a diverse…
Not Just the Destination, But the Journey: Reasoning Traces Causally Shape Generalization Behaviors
Pengcheng Wen, Yanxu Zhu, Jiapeng Sun +5
Chain-of-Thought (CoT) is often viewed as a window into LLM decision-making, yet recent work suggests it may function merely as post-hoc rationalization. This raises a critical ali…
AI Deception: Risks, Dynamics, and Controls
Boyuan Chen, Sitong Fang, Jiaming Ji +56
As intelligence increases, so does its shadow. AI deception, in which systems induce false beliefs to secure self-beneficial outcomes, has evolved from a speculative concern to an…
Reasoning Shapes Alignment: Investigating Cultural Alignment in Large Reasoning Models with Cultural Norms
Yuhang Wang, Yanxu Zhu, Jitao Sang
The advanced reasoning capabilities of Large Reasoning Models enable them to thoroughly understand and apply safety policies through deliberate thought processes, thereby improving…
Self-Guided Adaptive Safety Alignment: Synthesizing and Internalizing Guidelines in Reasoning Models
Yuhang Wang, Yanxu Zhu, Jiaming Zhang +2
Explicit safety policies can improve reasoning-model safety, but their effective coverage may lag behind evolving jailbreak strategies. We study whether a reasoning model can synth…
MoHoBench: Assessing Honesty of Multimodal Large Language Models via Unanswerable Visual Questions
Yanxu Zhu, Shitong Duan, Xiangxu Zhang +7
Recently Multimodal Large Language Models (MLLMs) have achieved considerable advancements in vision-language tasks, yet produce potentially harmful or untrustworthy content. Despit…
KG-FPQ: Evaluating Factuality Hallucination in LLMs with Knowledge Graph-based False Premise Questions
Yanxu Zhu, Jinlin Xiao, Yuhang Wang +1
Recent studies have demonstrated that large language models (LLMs) are susceptible to being misled by false premise questions (FPQs), leading to errors in factual knowledge, know a…
Improving Weak-to-Strong Generalization with Scalable Oversight and Ensemble Learning
Jitao Sang, Yuhang Wang, Jing Zhang +5
This paper presents a follow-up study to OpenAI's recent superalignment work on Weak-to-Strong Generalization (W2SG). Superalignment focuses on ensuring that high-level AI systems…
CDEval: A Benchmark for Measuring the Cultural Dimensions of Large Language Models
Yuhang Wang, Yanxu Zhu, Chao Kong +4
As the scaling of Large Language Models (LLMs) has dramatically enhanced their capabilities, there has been a growing focus on the alignment problem to ensure their responsible and…