Skip to content

Journal Club

InDeep PhD candidate Hosein Mohebbi organizes monthly online Zoom meetings with talks and discussion on recent research on interpretability. The Journal Club is especially meant for PhD students connected to the InDeep project, but everybody interested is welcome. Reach out to h.mohebbi@tilburguniversity.edu if you would like to join the talks.

DateSpeakerTopic/Paper
Jun 2024Martijn BentumTBA
May 2024Marcel Vélez Exploring the Inner Mechanisms of Large Generative Music Models
Apr 2024Jonathan KampThe Role of Syntactic Span Preferences in Post-Hoc Explanation Disagreement
Feb 2024Verna Dankers, University of EdinburghMemorisation in translation and classification: the what and the where
Nov 2023Ariana BisazzaCan modern LMs be truly polyglot?
Oct 2023Charlotte PouwExplaining Phone-Grapheme Mapping in Neural ASR models: A Case-Study on Place Assimilation
Sep 2023Jane Arleth dela CruzNoNE Found: Explaining the Output of Sequence-to-Sequence Models when No Named Entity is Recognized
May 2023John Ashley BurgoyneMusic Representation Learning
Apr 2023Tom LentzDiscovering Structure in Speech
Mar 2023Oskar van der Wal, University of AmsterdamThe Birth of Bias: A case study on the evolution of gender bias in an English language model
Mar 2023reading group led by Hosein MohebbiData distributional properties drive emergent in-context learning in transformers
Feb 2023reading group led by Michael HannaInterpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small
Nov 2022Sara Rajaee, University of AmsterdamThe Isotropy of Contextualized Spaces
Nov 2022reading group led by Jaap JumeletToy Models of Superposition
Apr 2022reading group led by Grzegorz ChrupałaLayer-wise Analysis of a Self-supervised Speech Representation Model
Feb 2022reading group led by Hosein MohebbiEvaluating Explanations: How much do explanations from the teacher aid students?
Feb 2022reading group led by Jonathan KampInterpretable Reinforcement Learning
Feb 2022reading group led by Gabriele SartiA Mathematical Framework for Transformer Circuits
Jan 2022reading group led by Hosein MohebbiIncorporating Residual and Normalization Layers into Analysis of Masked Language Models