Author

Zewen Liu

Observed name: Liu, Zewen

Institutions: Emory University, Qilu Institute of Technology

Observed institutions: Emory University, Qilu Institute of Technology

OpenAlex ID: A5137559621

Harness Updating Is Not Harness Benefit: Disentangling Evolution Capabilities in Self-Evolving LLM Agents

2026 · Minhua Lin, Juncheng Wu, et al. · Conference on Empirical Methods in Natural Language Processing (EMNLP) · 49

Adaptive Auto-Harness: Sustained Self-Improvement for Agentic System Deployment on Open-Ended Task Streams

2026 · Liu, Zewen, Shi, Zhan, et al. · arXiv (Cornell University) · 34

Evo-Harness: Context-to-Harness Skill Compilation for Self-Evolving Agents

2026 · Tianxin Wei, Zhan Shi, et al. · Conference on Empirical Methods in Natural Language Processing (EMNLP) · 26

Contagion Networks: Evaluator Preference Propagation in Multi-Agent LLM Systems

2026 · Liu, Zewen · arXiv (Cornell University) · 5

Memory Contagion: Cross-Temporal Propagation of Evaluator Bias via Agent Memory

2026 · Liu, Zewen · arXiv (Cornell University) · 2

BOUNDARY_SYNC: Measuring Communication-Induced Representational Coupling in Multi-Agent LLM Systems

2026 · Liu, Zewen · arXiv (Cornell University) · 1

Calibrating the Evaluator: Does Probability Calibration Mitigate Preference Coupling in LLM Agent Feedback Loops?

2026 · Liu, Zewen · arXiv (Cornell University) · 1

The Hidden Cost of Resampling: How Imbalance Correction Degrades Probability Calibration in Tree Ensembles

2026 · Liu, Zewen · arXiv (Cornell University) · 1

EPC: A Standardized Protocol for Measuring Evaluator Preference Dynamics in LLM Agent Systems

2026 · Liu, Zewen · arXiv (Cornell University)

The Contagion Tensor: A Framework for Measuring Output-Distribution Coupling in Multi-Agent LLM Systems -- and Auditing the Claims It Enables

2026 · Liu, Zewen · arXiv (Cornell University)

Works count: 10

Citation count: 119

H-index: 4

I10-index: 3