That's Deprecated! Understanding, Detecting, and Steering Knowledge Conflicts in Language Models for Code Generation

Jaesung Bae; Cameron Churchwell; Mitchell Hermon; Tsun-An Hsieh; Jocelyn Xu; Yekaterina Yegorova; Mark Hasegawa-Johnson; Heng Ji

arXiv, 2025

That's Deprecated! Understanding, Detecting, and Steering Knowledge Conflicts in Language Models for Code Generation

Jaesung Bae, Cameron Churchwell, Mitchell Hermon, Tsun-An Hsieh, Jocelyn Xu, Yekaterina Yegorova, Mark Hasegawa-Johnson, Heng Ji

A study of how language models handle conflicts between learned programming knowledge and prompt-provided updates, with probes and activation steering for code generation.

LLM BehaviorCode GenerationKnowledge Conflicts

Project page Paper

This work studies how language models respond when a prompt conflicts with their learned programming knowledge. It extends knowledge-conflict analysis to code generation and evaluates whether those conflicts can be detected and steered from model activations.