Workshop Schedule

Legend: contributed · invited · break
Time Program Format
09:00 - 09:10 Introduction In-person
09:10 - 9:40 Invited Talk: Causal effects in the eyes of the beholder Francesco Locatello
9:40 - 10:10 Invited Talk: Learning Programming Abstractions for Automating Formal Verification Qiaochu Chen
10:10 - 10:30 Coffee break -
10:30 - 10:45 Navigating the Concept Space of Language Models Wilson Marcílio
10:45 - 11:00 Diagnosing and Fixing Latent Recovery in Sparse Autoencoders Marco Fumero
11:00 - 12:00 Poster session In-person
12:00 - 13:00 Lunch break -
13:00 - 13:30 Invited Talk: Learning Causal Systems in the Latent Space Biwei Huang
13:30 - 14:00 Invited Talk: In Defense of Concept Leakage Mateo Espinosa Zarlenga
14:00 - 14:15 Automated Concept Discovery for LLM-as-a-Judge Preference Analysis James Wedgwood
14:15 - 14:45 Invited Talk: Humans & LLMs: Two Great Systems Separated by a Common Language Subbarao Kamhampati
14:45 - 15:15 Invited Talk: Expanding our concepts with help of agents Been Kim
15:15 - 15:30 Coffee break In-person
15:30 - 15:45 Dynamics Reveals Structure: Challenging the Linear Propagation Assumption Hoyeon Chang
15:45 - 16:00 Automatic Interpretation of Visual Concepts Felix Meissen
16:00 - 17:00 Poster session In-person
All times according to UCT-3.

Poster Session

All accepted papers will be presented as a poster at the workshop. All posters can be put up in the morning and can remain up during the whole day (i.e., all posters are presented during both poster sessions).