The key takeaway of this study is that compared to the human cloze baseline, language models over-predict words that are either related to the most predictable next word (the 'best completion') or to the event under discussion

Really enjoyed the babyLM talks and posters hosted by CoNLL 2024/Cognitive Modeling and Computational Linguistics at EMNLP 2024 last year! Looking forward to seeing what people come up with this time round!

This is concerning, and I wouldn't be surprised if it leads to some students having to withdraw their papers from the conference

Exciting to see our paper (with Megan Bardolph, Cyma K. Van Petten, Benjamin K. Bergen, and Seana Coulson) 'in print' at Neurobiology of Language!

This paper is now accepted to be presented at Conference on Language Modeling! Updated version is on arXiv. Feeling excited for the conference, let me know if you want to meet!