从行为表现到内在能力：用VLM-Lens解读视觉-语言模型
From Behavioral Performance to Internal Competence: Interpreting Vision-Language Models with VLM-Lens
AI总结：
本研究推出开源工具包VLM-Lens，支持从16款主流基础VLM及其30余个变体的任意层提取中间输出，提供统一配置接口屏蔽模型差异，可适配新模型，助力VLM的可解释性分析与性能改进。
摘要：
我们推出VLM-Lens，这是一款旨在实现视觉-语言模型（VLM）系统性基准测试、分析与解读的工具包，支持从开源VLM前向传播过程中的任意层提取中间输出。VLM-Lens提供统一的、支持YAML配置的接口，屏蔽了不同模型的特定复杂度，可在各类VLM上实现用户友好的操作。它目前支持16个最先进的基础VLM及其30余个变体，且具备可扩展性，无需修改核心逻辑即可适配新模型。\n该工具包可轻松与各类可解释性分析方法集成。我们通过两项简单的分析实验展示了其用法，揭示了VLM不同层间、针对不同目标概念的隐藏表示存在系统性差异。VLM-Lens作为开源项目发布，旨在推动社区在理解与改进VLM方面的工作。
英文摘要：
We introduce VLM-Lens, a toolkit designed to enable systematic benchmarking, analysis, and interpretation of vision-language models (VLMs) by supporting the extraction of intermediate outputs from any layer during the forward pass of open-source VLMs. VLM-Lens provides a unified, YAML-configurable interface that abstracts away model-specific complexities and supports user-friendly operation across diverse VLMs. It currently supports 16 state-of-the-art base VLMs and their over 30 variants, and is extensible to accommodate new models without changing the core logic. The toolkit integrates easily with various interpretability and analysis methods. We demonstrate its usage with two simple analytical experiments, revealing systematic differences in the hidden representations of VLMs across layers and target concepts. VLM-Lens is released as an open-sourced project to accelerate community efforts in understanding and improving VLMs.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
