一作ACL 2026Findings· Findings of ACL
Preserving Language Capabilities in Vision-Language Models via Representation Regulation
ZiXuan Chen · Juncheng Tao · Ziqian Zeng
摘要
Vision-Language Models (VLMs) provide a unified framework to process both text-only tasks and vision-language tasks. However, finetuning VLMs on vision-language data has degraded language capabilities. In this paper, we prove that as the training loss declines during finetuning, the visual repres...