Zekun Wu

36 papers A* 2B 2Journal 24Unranked 8
YearRankTypeTitle / Venue / Authors
2026 J jnl
AgentDrift: Unsafe Recommendation Drift Under Tool Corruption Hidden by Ranking Metrics in LLM Agents.
CoRR
Zekun Wu, Adriano S. Koshiyama, Sahan Bulathwela, María Pérez-Ortiz
2026 A* conf
AgentGraph: Trace-to-Graph Platform for Interactive Analysis and Robustness Testing in Agentic AI Systems.
AAAI
Zekun Wu, Seonglae Cho, Cristian E. Muñoz Villalobos, Theo King, Umar Mohammed, Emre Kazim, María Pérez-Ortiz, Sahan Bulathwela, Adriano S. Koshiyama
2026 A* conf
AgentSeer: Visualizing and Evaluating Temporal Actions in Agentic AI Systems.
AAAI
Ilham Wicaksono, Zekun Wu, Rahul Patel, Theo King, Adriano S. Koshiyama, Philip Colin Treleaven
2026 J jnl
Control Reinforcement Learning: Interpretable Token-Level Steering of LLMs via Sparse Autoencoder Features.
CoRR
Seonglae Cho, Zekun Wu, Adriano S. Koshiyama
2026 J jnl
The Confidence Manifold: Geometric Structure of Correctness Representations in Language Models.
CoRR
Seonglae Cho, Zekun Wu, Kleyton da Costa, Adriano S. Koshiyama
2025 — conf
Bias Amplification: Large Language Models as Increasingly Biased Media.
IJCNLP-AACL (long papers)
Ze Wang, Zekun Wu, Yichi Zhang, Xin Guan, Navya Jain, Qinyang Lu, Saloni Gupta, Adriano S. Koshiyama
2025 B conf
CauSkelNet: Causal Representation Learning for Human Behaviour Analysis.
FG
Xingrui Gu, Chuyi Jiang, Erte Wang, Zekun Wu, Qiang Cui, Leimin Tian, Lianlong Wu, Siyang Song, Chuang Yu
2025 J jnl
CorrSteer: Steering Improves Task Performance and Safety in LLMs through Correlation-based Sparse Autoencoder Feature Selection.
CoRR
Seonglae Cho, Zekun Wu, Adriano S. Koshiyama
2025 — conf
From Text to Emoji: How PEFT-Driven Personality Manipulation Unleashes the Emoji Potential in LLMs.
NAACL (Findings)
Navya Jain, Zekun Wu, Cristian E. Muñoz Villalobos, Airlie Hilliard, Xin Guan, Adriano S. Koshiyama, Emre Kazim, Philip C. Treleaven
2025 — conf
HEARTS: A Holistic Framework for Explainable, Sustainable and Robust Text Stereotype Detection.
IJCNLP-AACL (long papers)
Theo King, Zekun Wu, Adriano S. Koshiyama, Emre Kazim, Philip C. Treleaven
2025 — conf
HyPA-RAG: A Hybrid Parameter Adaptive Retrieval-Augmented Generation System for AI Legal and Policy Applications.
NAACL (Industry Track)
Rishi Kalra, Zekun Wu, Ayesha Gulley, Airlie Hilliard, Xin Guan, Adriano S. Koshiyama, Philip C. Treleaven
2025 J jnl
Knowledge Collapse in LLMs: When Fluency Survives but Facts Fail under Recursive Synthetic Training.
CoRR
Figarri Keisha, Zekun Wu, Ze Wang, Adriano S. Koshiyama, Philip C. Treleaven
2025 — conf
LibVulnWatch: A Deep Assessment Agent System and Leaderboard for Uncovering Hidden Vulnerabilities in Open-Source AI Libraries.
ACL (4)
Zekun Wu, Seonglae Cho, Umar Mohammed, Cristian E. Muñoz Villalobos, Kleyton da Costa, Xin Guan, Theo King, Ze Wang, Emre Kazim, Adriano S. Koshiyama
2025 J jnl
LibVulnWatch: A Deep Assessment Agent System and Leaderboard for Uncovering Hidden Vulnerabilities in Open-Source AI Libraries.
CoRR
Zekun Wu, Seonglae Cho, Umar Mohammed, Cristian E. Muñoz, Kleyton da Costa, Xin Guan, Theo King, Ze Wang, Emre Kazim, Adriano S. Koshiyama
2025 J jnl
MPF: Aligning and Debiasing Language Models post Deployment via Multi Perspective Fusion.
CoRR
Xin Guan, PeiHsin Lin, Zekun Wu, Ze Wang, Ruibo Zhang, Emre Kazim, Adriano S. Koshiyama
2025 — conf
MPF: Aligning and Debiasing Language Models post Deployment via Multi-Perspective Fusion.
IJCNLP-AACL (Findings)
Xin Guan, PeiHsin Lin, Zekun Wu, Ze Wang, Ruibo Zhang, Emre Kazim, Adriano S. Koshiyama
2025 J jnl
Mind the Gap: Comparing Model- vs Agentic-Level Red Teaming with Action-Graph Observability on GPT-OSS-20B.
CoRR
Ilham Wicaksono, Zekun Wu, Rahul Patel, Theo King, Adriano S. Koshiyama, Philip C. Treleaven
2025 J jnl
Mind the Gap: Evaluating Model- and Agentic-Level Vulnerabilities in LLMs with Action Graphs.
CoRR
Ilham Wicaksono, Zekun Wu, Rahul Patel, Theo King, Adriano S. Koshiyama, Philip C. Treleaven
2025 J jnl
NormCode: A Semi-Formal Language for Auditable AI Planning.
CoRR
Xin Guan, Yunshan Li, Zekun Wu, Ruibo Zhang
2025 J jnl
Personality as a Probe for LLM Evaluation: Method Trade-offs and Downstream Effects.
CoRR
Gunmay Handa, Zekun Wu, Adriano S. Koshiyama, Philip C. Treleaven
2025 B conf
SAGED: A Holistic Bias-Benchmarking Pipeline for Language Models with Customisable Fairness Calibration.
COLING
Xin Guan, Nathaniel Demchak, Saloni Gupta, Ze Wang, Ediz Ertekin Jr., Adriano S. Koshiyama, Emre Kazim, Zekun Wu
2024 J jnl
Advancing Multimodal Data Fusion in Pain Recognition: A Strategy Leveraging Statistical Correlation and Human-Centered Perspectives.
CoRR
Xingrui Gu, Zhixuan Wang, Irisa Jin, Zekun Wu
2024 — conf
Advancing Pain Recognition Through Statistical Correlation-Driven Multimodal Fusion.
ACII (Workshops and Demos)
Xingrui Gu, Zhixuan Wang, Irisa Jin, Zekun Wu
2024 J jnl
Assessing Bias in Metric Models for LLM Open-Ended Generation Bias Benchmarks.
CoRR
Nathaniel Demchak, Xin Guan, Zekun Wu, Ziyi Xu, Adriano S. Koshiyama, Emre Kazim
2024 J jnl
Auditing Large Language Models for Enhanced Text-Based Stereotype Detection and Probing-Based Bias Evaluation.
CoRR
Zekun Wu, Sahan Bulathwela, María Pérez-Ortiz, Adriano Soares Koshiyama
2024 J jnl
Bias Amplification: Language Models as Increasingly Biased Media.
CoRR
Ze Wang, Zekun Wu, Jeremy Zhang, Navya Jain, Xin Guan, Adriano S. Koshiyama
2024 J jnl
CauSkelNet: Causal Representation Learning for Human Behaviour Analysis.
CoRR
Xingrui Gu, Chuyi Jiang, Erte Wang, Zekun Wu, Qiang Cui, Leimin Tian, Lianlong Wu, Siyang Song, Chuang Yu
2024 J jnl
Eliciting Personality Traits in Large Language Models.
CoRR
Airlie Hilliard, Cristian E. Muñoz, Zekun Wu, Adriano Soares Koshiyama
2024 J jnl
From Text to Emoji: How PEFT-Driven Personality Manipulation Unleashes the Emoji Potential in LLMs.
CoRR
Navya Jain, Zekun Wu, Cristian E. Muñoz, Airlie Hilliard, Adriano S. Koshiyama, Emre Kazim, Philip C. Treleaven
2024 J jnl
HEARTS: A Holistic Framework for Explainable, Sustainable and Robust Text Stereotype Detection.
CoRR
Theo King, Zekun Wu, Adriano S. Koshiyama, Emre Kazim, Philip C. Treleaven
2024 J jnl
HyPA-RAG: A Hybrid Parameter Adaptive Retrieval-Augmented Generation System for AI Legal and Policy Applications.
CoRR
Rishi Kalra, Zekun Wu, Ayesha Gulley, Airlie Hilliard, Xin Guan, Adriano S. Koshiyama, Philip C. Treleaven
2024 — conf
JobFair: A Framework for Benchmarking Gender Hiring Bias in Large Language Models.
EMNLP (Findings)
Ze Wang, Zekun Wu, Xin Guan, Michael Thaler, Adriano S. Koshiyama, Skylar Lu, Sachin Beepath, Ediz Ertekin Jr., María Pérez-Ortiz
2024 J jnl
JobFair: A Framework for Benchmarking Gender Hiring Bias in Large Language Models.
CoRR
Ze Wang, Zekun Wu, Xin Guan, Michael Thaler, Adriano S. Koshiyama, Skylar Lu, Sachin Beepath, Ediz Ertekin Jr., María Pérez-Ortiz
2024 J jnl
SAGED: A Holistic Bias-Benchmarking Pipeline for Language Models with Customisable Fairness Calibration.
CoRR
Xin Guan, Nathaniel Demchak, Saloni Gupta, Ze Wang, Ediz Ertekin Jr., Adriano S. Koshiyama, Emre Kazim, Zekun Wu
2024 J jnl
THaMES: An End-to-End Tool for Hallucination Mitigation and Evaluation in Large Language Models.
CoRR
Mengfei Liang, Archish Arun, Zekun Wu, Cristian E. Muñoz, Jonathan Lutch, Emre Kazim, Adriano S. Koshiyama, Philip C. Treleaven
2023 J jnl
Towards Auditing Large Language Models: Improving Text-based Stereotype Detection.
CoRR
Zekun Wu, Sahan Bulathwela, Adriano Soares Koshiyama