---
title: VLM-Lens
emoji: 👀
colorFrom: blue
colorTo: indigo
sdk: gradio
sdk_version: 5.48.0
app_file: app.py
pinned: false
license: apache-2.0
thumbnail: >-
  /static-proxy?url=https%3A%2F%2Fcdn-uploads.huggingface.co%2Fproduction%2Fuploads%2F630cfc45b66f088d547b2768%2Ff3VCIcopD2bzyP2XdRa1T.png
short_description: '[EMNLP 2025 Demo] VLM-Lens: Extracting VLM representations'
---

# VLM-Lens 👁️🔍

A visual lens into the internals of Vision-Language Models.  
Built with Gradio, this demo lets you explore token-level probabilities, spatial grounding, and interpretability visualizations.

> Developed by [@marstin](https://huggingface.co/marstin)