AI safety researcher

Xiaowei Huang

University of Liverpool

Google Scholar dblp OpenReview ORCID Homepage

Last 12 months12 papers

Attacks 2

AI safety papers from the last year

On this site: 71 papers

October 2026

The One-Pixel of Reasoning: One Truthful Relation Is Enough to Mislead LLM Agents
Oct 17AttacksarXiv ↗
Organising Trajectory Evidence for Language-Model Agent Assurance: Fragments, Methods, and the Residual
Oct 3arXiv ↗

August 2026

SPK: Eliciting Structured Prior Knowledge for Interpretable Out-of-Distribution Detection in Real-Time Object Detection
Aug 19arXiv ↗

July 2026

SkillFuzz: Fuzzing Skill Composition for Implicit Intents Discovery in Open Skill Marketplaces
Jul 2arXiv ↗

June 2026

SCARCE: Scalable Cascade Analysis for Rare-event Characterisation via Embeddings
Jun 28arXiv ↗

May 2026

March 2026

S2O: Enhancing Adversarial Training with Second-Order Statistics of Weights
Mar 1arXiv ↗

January 2026

Lying with Truths: Open-Channel Multi-Agent Collusion for Belief Manipulation via Generative Montage
Jan 4AttacksarXiv ↗

November 2025

Fragile by Design: On the Limits of Adversarial Defenses in Personalized Generation
Nov 13arXiv ↗

Authorship follows dblp and the author's own Google Scholar profile; only AI safety papers on this site are listed.