9:00-9:10 AM
Opening Remarks
9:10-10:00 AM
Tutorial: Sonia Joseph
Prisma: An Open Source Toolkit for Mechanistic Interpretability in Vision and Video
10:00-10:20 AM
Coffee Break
11:20 AM-12:00 PM
Spotlight Talks (3 x 10 min) + 10 min Q&A:
Investigating Mechanisms for In-Context Vision Language Binding
Analyzing Hierarchical Structure in Vision Models with Sparse Autoencoders
Disentangling Polysemantic Channels in Convolutional Neural Networks
12:00-12:30 PM
Invited Talk: Aleksander Madry
12:30-2:30 PM
Lunch Break
12:30-2:30 PM
Poster Session (ExHall D, Posters #31 - #60)
2:30-3:10 PM
Spotlight Talks (3 x 10 min) + 10 min Q&A:
Too Late to Recall: The Two-Hop Problem in Multimodal Knowledge Retrieval
Line of Sight: On Linear Representations in VLLMs
Revelio: Interpreting and leveraging semantic information in diffusion models
3:10-3:30 PM
Coffee Break
4:30-4:35 PM
Closing Remarks