Publications

Conference Papers

Focus-LIME: Surgical Interpretation of Long-Context Large Language Models via Proxy-Based Neighborhood Selection

Published in IJCAI-ECAI 2026, 2026

A coarse-to-fine framework that uses a proxy model to select an optimized neighborhood for faithful, fine-grained explanations of long-context large language models.

Recommended citation: Junhao Liu, Haonan Yu, Zhenyu Yan, and Xin Zhang. (2026). "Focus-LIME: Surgical Interpretation of Long-Context Large Language Models via Proxy-Based Neighborhood Selection." IJCAI-ECAI 2026.
Download Paper

MAnchors: Memorization-Based Acceleration of Anchors via Rule Reuse and Transformation

Published in International Conference on Machine Learning (ICML), 2026

A memorization-based framework that accelerates Anchors by reusing and transforming explanation rules while preserving fidelity and understandability.

Recommended citation: Haonan Yu, Junhao Liu and Xin Zhang et al. (2026). "MAnchors: Memorization-Based Acceleration of Anchors via Rule Reuse and Transformation." International Conference on Machine Learning (ICML).
Download Paper

Revitalizing Black-Box Interpretability: Actionable Interpretability for LLMs via Proxy Models

Published in Annual Meeting of the Association for Computational Linguistics (ACL), 2026

A budget-friendly proxy framework for generating faithful model-agnostic explanations for expensive black-box LLMs.

Recommended citation: Junhao Liu, Haonan Yu, Zhenyu Yan, and Xin Zhang. (2026). "Revitalizing Black-Box Interpretability: Actionable Interpretability for LLMs via Proxy Models." Annual Meeting of the Association for Computational Linguistics (ACL).
Download Paper