Yihang Chen
10 indexed papers
Publications per year
Top categories
Frequent co-authors
Research Timeline
The paper proposes AnaFP, a theoretically guided analytical fingerprinting scheme that determines the optimal distance of a model's fingerprint from the decision boundary to ensure both robustness and uniqueness for model ownership protection.
IrisFP introduces a novel adversarial-example-based framework that generates composite-sample fingerprints near the intersection of multiple decision boundaries, significantly enhancing model ownership verification robustness and uniqueness.
LiteGuard proposes an efficient task-agnostic model fingerprinting framework that achieves enhanced generalization and significantly reduces computational overhead compared to existing methods like MetaV.
The paper introduces FORCEBENCH, a new stress test designed to evaluate whether cited sources genuinely warrant the strength of a claim, revealing that standard citation evaluation methods often fail to detect over-strong claims.
The paper introduces FORCEBENCH, a new stress test designed to evaluate whether cited sources genuinely warrant the strength of a claim, revealing that standard citation evaluation methods often fail to detect over-strong claims.
The paper introduces SkillReact, a framework that measures compositional risk in agent skill ecosystems, finding that even if individual skills are safe, their combination can create significant, exploitable security vulnerabilities.
The paper introduces SkillReact, a framework that measures compositional risk in agent skill ecosystems, finding that even if individual skills are safe, their combination can create significant, exploitable security vulnerabilities.
The paper introduces SkillReact, a framework that measures compositional risk in agent skill ecosystems, finding that even if individual skills are safe, their combination can create significant, unaddressed security vulnerabilities.
This paper reinterprets supervised fine-tuning as target distribution design and proposes a new method called Target-SFT that constructs the training objective directly from the desired target distribution.
This paper separates three mechanisms contributing to pipeline safety in multi-agent large language models and introduces a controlled contrast design to evaluate their impact.
Papers
Operational Reframing and Approval-Framed Delegation in Multi-Agent LLM Safety
Lifei Liu, Haoran Yu, Xiaochong Jiang, Su Wang +2 more
This paper separates three mechanisms contributing to pipeline safety in multi-agent large language models and introduces a controlled contrast design to evaluate their impact.