Publications

Surface Compliance paper figure ACL 2026 Findings

The Model Agreed, But Didn’t Learn: Diagnosing Surface Compliance in Large Language Models

Gu, X.*, Huang, Z.*, Hong, W., Xie, J., Lou, R., & Zhang, K.

We propose SA-MCQ, a diagnostic framework that reveals surface compliance in knowledge editing, where models appear to accept edits without truly internalizing them.

UltraEdit paper figure TMLR 2026

UltraEdit: Training-, Subject-, and Memory-Free Lifelong Editing in Language Models

Gu, X., Huang, Z., Gu, J.-C., & Zhang, K.

UltraEdit is a lifelong editing framework for language models that removes the need for extra training, subject-specific memory, and memory storage.