Mechanistic Interpretability for Vision

Workshop

Mechanistic Interpretability for Vision

Tamar Rott Shaham · Yossi Gandelsman · Joanna Materzynska · Rohit Gandikota · Amil Dravid · Ashkan Khakzar · Eli Shechtman · Philip H.S. Torr

Thu 12 Jun 7 a.m. PDT — 3:30 p.m. PDT

[ Abstract ] Workshop Website

[ Project Page ]

Though start and end times here are correct, detailed schedules here may not be complete or up to date. Please be sure to cross reference the workshop's website to verify workshop schedule details if they are available on the workshop's website. (Added by CVPR.)

Chat is not available.

Timezone: America/Los_Angeles

Schedule

-	Disentangling Polysemantic Channels in Convolutional Neural Networks ( Paper ) > link Link	Robin Hesse · Jonas Fischer · Simone Schaub-Meyer · Stefan Roth 🔗
-	Naturally Computed Scale Invariance in the Residual Stream of ResNet18 ( Paper ) >	André Longon 🔗
-	Decoding Vision Transformers: the Diffusion Steering Lens ( Paper ) >	Ryota Takatsuki · Sonia Joseph · Ippei Fujisawa · Ryota Kanai 🔗
-	Visualizing and Controlling Cortical Responses Using Voxel-Weighted Activation Maximization ( Paper ) >	Matthew Shinkle · Mark Lescroart 🔗
-	Analyzing Hierarchical Structure in Vision Models with Sparse Autoencoders ( Paper ) >	Matthew L. Olson · Musashi Hinck · Neale Ratzlaff · Changbai Li · Phillip Howard · Vasudev Lal · Shao-Yen Tseng 🔗
-	Leveraging Vision-Language Foundation Models to Reveal Hidden Image-Attribute Relationships in Medical Imaging ( Paper ) >	Amar Kumar · Anita Kriz · Barak Pertzov · Tal Arbel 🔗
-	Language-Guided Trajectory Traversal in Disentangled Stable Diffusion Latent Space for Factorized Medical Image Generation ( Paper ) >	Zahra TehraniNasab · Amar Kumar · Tal Arbel 🔗
-	Investigating Mechanisms for In-Context Vision Language Binding ( Paper ) >	Darshana Saravanan · Makarand Tapaswi · Vineet Gandhi 🔗
-	Uncovering Branch-specialization in InceptionV1 using k sparse autoencoders ( Paper ) >	Matthew Bozoukov Matthew Bozoukov 🔗
-	Embedding Shift Dissection on CLIP: Effects of Augmentations on VLM’s Representation Learning ( Paper ) >	Ashim Dahal · Saydul Murad Murad · Nick Rahimi 🔗
-	Wavelet-Based Mechanistic Interpretability of Vision Transformers via Frequency-Aware Ablations ( Paper ) >	Sophia Abraham · Jonathan Hauenstein · Walter Scheirer 🔗