一作ACL 2026Workshop· ACL System Demonstrations
Spectra: A Mechanistic Interpretability Library for Vision-Language Models
Clement Neo · Yongsen Zheng · Kwok Yan Lam · Luke Ong
摘要
Vision-Language Models (VLMs) have become increasingly important in AI applications, yet interpretability tools for these models lag behind those available for text-only language models. While libraries like TransformerLens have enabled significant progress in understanding language models, exist...