roxpetcu.bsky.social
Mechanistic Interpretability for Retrieval and Reasoning. PhD student at IRLab Amsterdam 🌧️
👤 Joined on 18 November 2024
Followers
Following
Total Posts
@roxpetcu.bsky.social_matatini_ Mata🤘 "Here’s a look back at a timeless moment when a curious whale calf danced in front of us. When Mother Nature chooses you, everything becomes different...⏳🔮"
Show PostSentences are partially understood before they're fully read. How do LMs incrementally interpret their inputs? In a new paper, @amuuueller.bsky.social and I use mech interp tools to study how LMs process structurally ambiguous sentences. We show LMs rely on both syntactic & spurious features! 1/10
Show PostCome see @eijkelboomfloor.bsky.social and @gbarto.bsky.social present their work on variational flow matching now in West Ballroom A-D #7103!
Show Post