4 citations · 5 across the 19 of their papers we have counts for
22 papers
EgoSAT: A Comprehensive Benchmark of Egocentric Streaming Interaction Understanding
Yijia Lei, Jinzhao Li, Yichi Zhang +3
We introduce EgoSAT, the first comprehensive benchmark for egocentric video reasoning in streaming settings, designed to evaluate the capabilities of modern vision-language models…
MESA: Improving MoE Safety Alignment via Decentralized Expertise
Yitong Sun, Yao Huang, Teng Li +5
Mixture-of-Experts (MoE) architectures scale Large Language Models (LLMs) efficiently, enabling greater capacity with reduced computational cost by dynamically routing inputs to re…
IPIBench: Evaluating Interactive Proactive Intelligence of MLLMs under Continuous Streams
Jinzhao Li, Yinuo Chen, Wenxuan Song +5
Recent multimodal large language models (MLLMs) achieve strong performance on reactive question answering, but real-world streaming assistants require proactive reasoning over cont…
RepoMirage: Probing Repository Context Reasoning in Code Agents with Perturbations
Hanyu Li, Yichi Zhang, Speed Zhu +3
Code agents are currently having skillful performance on repository-level software engineering benchmarks, but it remains unclear whether success on end-to-end tasks such as issue…
Guideline-Grounded Evidence Accumulation for High-Stakes Agent Verification
Yichi Zhang, Nabeel Seedat, Yinpeng Dong +3
As LLM-powered agents have been used for high-stakes decision-making, such as clinical diagnosis, it becomes critical to develop reliable verification of their decisions to facilit…
Reasoning as State Transition: A Representational Analysis of Reasoning Evolution in Large Language Models
Siyuan Zhang, Jialian Li, Yichi Zhang +3
Large Language Models have achieved remarkable performance on reasoning tasks, motivating research into how this ability evolves during training. Prior work has primarily analyzed…
DeceptionBench: A Comprehensive Benchmark for AI Deception Behaviors in Real-world Scenarios
Yao Huang, Yitong Sun, Yichi Zhang +3
Despite the remarkable advances of Large Language Models (LLMs) across diverse cognitive tasks, the rapid enhancement of these capabilities also introduces emergent deceptive behav…
Oyster-I: Beyond Refusal -- Constructive Safety Alignment for Responsible Language Models
Ranjie Duan, Jiexi Liu, Xiaojun Jia +27
Large language models (LLMs) typically deploy safety mechanisms to prevent harmful content generation. Most current approaches focus narrowly on risks posed by malicious actors, of…
Towards Safe Reasoning in Large Reasoning Models via Corrective Intervention
Yichi Zhang, Yue Ding, Jingwen Yang +7
Although Large Reasoning Models (LRMs) have progressed in solving complex problems, their chain-of-thought (CoT) reasoning often contains harmful content that can persist even when…
Unveiling Trust in Multimodal Large Language Models: Evaluation, Analysis, and Mitigation
Yichi Zhang, Yao Huang, Yifan Wang +10
The trustworthiness of Multimodal Large Language Models (MLLMs) remains an intense concern despite the significant progress in their capabilities. Existing evaluation and mitigatio…
A Survey on Autonomy-Induced Security Risks in Large Model-Based Agents
Hang Su, Jun Luo, Chang Liu +4
Recent advances in large language models (LLMs) have catalyzed the rise of autonomous AI agents capable of perceiving, reasoning, and acting in dynamic, open-ended environments. Th…
Breaking the Ceiling: Exploring the Potential of Jailbreak Attacks through Expanding Strategy Space
Yao Huang, Yitong Sun, Shouwei Ruan +3
Large Language Models (LLMs), despite advanced general capabilities, still suffer from numerous safety risks, especially jailbreak attacks that bypass safety protocols. Understandi…
Beyond Quantity: Distribution-Aware Labeling for Visual Grounding
Yichi Zhang, Gongwei Chen, Jun Zhu +2
Visual grounding requires large and diverse region-text pairs. However, manual annotation is costly and fixed vocabularies restrict scalability and generalization. Existing pseudo-…
Mitigating Overthinking in Large Reasoning Models via Manifold Steering
Yao Huang, Huanran Chen, Shouwei Ruan +3
Recent advances in Large Reasoning Models (LRMs) have demonstrated remarkable capabilities in solving complex tasks such as mathematics and coding. However, these models frequently…
Unveiling the Basin-Like Loss Landscape in Large Language Models
Huanran Chen, Yinpeng Dong, Zeming Wei +4
We discover the emergence of \textit{basins} in the loss landscape of large language models. As model scale increases, LLMs become progressively more resilient to random perturbati…
RealSafe-R1: Safety-Aligned DeepSeek-R1 without Compromising Reasoning Capability
Yichi Zhang, Zihao Zeng, Dongbai Li +3
Large Reasoning Models (LRMs), such as OpenAI o1 and DeepSeek-R1, have been rapidly progressing and achieving breakthrough performance on complex reasoning tasks such as mathematic…
Exploring the Generalizability of Factual Hallucination Mitigation via Enhancing Precise Knowledge Utilization
Siyuan Zhang, Yichi Zhang, Yinpeng Dong +1
Large Language Models (LLMs) often struggle to align their responses with objective facts, resulting in the issue of factual hallucinations, which can be difficult to detect and mi…
STAIR: Improving Safety Alignment with Introspective Reasoning
Yichi Zhang, Siyuan Zhang, Yao Huang +7
Ensuring the safety and harmlessness of Large Language Models (LLMs) has become equally critical as their performance in applications. However, existing safety alignment methods ty…
Scaling Laws for Black box Adversarial Attacks
Chuan Liu, Huanran Chen, Yichi Zhang +2
Adversarial examples exhibit cross-model transferability, enabling threatening black-box attacks on commercial models. Model ensembling, which attacks multiple surrogate models, is…
MultiTrust: A Comprehensive Benchmark Towards Trustworthy Multimodal Large Language Models
Yichi Zhang, Yao Huang, Yitong Sun +10
Despite the superior capabilities of Multimodal Large Language Models (MLLMs) across diverse tasks, they still face significant trustworthiness challenges. Yet, current literature…
Exploring the Transferability of Visual Prompting for Multimodal Large Language Models
Yichi Zhang, Yinpeng Dong, Siyuan Zhang +3
Although Multimodal Large Language Models (MLLMs) have demonstrated promising versatile capabilities, their performance is still inferior to specialized models on downstream tasks,…
How Robust is Google's Bard to Adversarial Image Attacks?
Yinpeng Dong, Huanran Chen, Jiawei Chen +6
Multimodal Large Language Models (MLLMs) that integrate text and other modalities (especially vision) have achieved unprecedented performance in various multimodal tasks. However,…