Postdoc @UniofOxford researching language in humans and machines. Previously: @MIT @CogSciUCSD, @InfAtEd, @SchoolofPPLS
Joined September 2017
- Tweets89
- Following799
- Followers422
- Likes368
Looking forward to #EMNLP2026! I’ll be presenting our recent work on applying activation steering methods to generating text in a specific target language at Poster Session F
We show that an optimal control method (LiSeCo) generally outperforms the standard difference-in-means approach while requiring substantially less hyperparameter tuning
See the full paper here: arxiv.org/abs/2609.16967
Looking forward to #SNL2026! I’ll be presenting the work published in our recent JML paper about language model scaling and the N400. Find me (or email/message) if you want to chat about predictive coding and NLP in the study of human language comprehension! Paper below:
I’ll be in San Diego for #ACL2026 and presenting this work at @conll_conf #CoNLL2026 on July 3rd! Feel free to reach out if you want to chat!
Excited to announce that I’ll be presenting a paper at #NeurIPS this year! Reach out if you’re interested in chatting about LM training dynamics, architectural differences, shortcuts/heuristics, or anything at the CogSci/NLP/AI interface in general! #Neurips2025
Seems like a good time to share our new preprint about model openness! (with @linguist_cat @tylerachang @PamRiv1 @SamuelTaylorCS @camrobjones @Sean_Trott @roger_p_levy Ben Bergen @drmaltman):
We also discuss other nuances, including factors to consider in safety, socio-technical, and HCI research; approaches to mitigating the problems associated with closed-weight models; and the limits of what open weights alone can provide
Here’s a summary of our main conclusions, and a link to the paper: arxiv.org/abs/2603.26539
James Michaelov @ EMNLP retweeted
In film, "we'll fix it in post" is what you say when something went wrong on set and you don't want to redo it. AI research has made it our entire methodology: train the model, then patch whatever comes out. Our new ICML oral argues this can't be the basis of a science of AI. 🧵
Had a great first day at #HSP2026 yesterday! Looking forward to presenting on the relationship between reading time, n-grams, and language model scaling at the 12.10-2pm poster session today!
Presenting this at the poster session this morning (11-2pm) at #5109
Excited to announce that I’ll be presenting a paper at #NeurIPS this year! Reach out if you’re interested in chatting about LM training dynamics, architectural differences, shortcuts/heuristics, or anything at the CogSci/NLP/AI interface in general! #Neurips2025
Looking forward to #NeurIPS25 this week 🏝️! I'll be presenting at Poster Session 3 (11-2 on Thursday). Feel free to reach out!
Excited to announce that I’ll be presenting a paper at #NeurIPS this year! Reach out if you’re interested in chatting about LM training dynamics, architectural differences, shortcuts/heuristics, or anything at the CogSci/NLP/AI interface in general! #Neurips2025
Excited to announce that I’ll be presenting a paper at #NeurIPS this year! Reach out if you’re interested in chatting about LM training dynamics, architectural differences, shortcuts/heuristics, or anything at the CogSci/NLP/AI interface in general! #Neurips2025
I'll also be presenting this paper with @linguist_cat at #CogInterp! xitter.cf/linguist_cat/status/19…
@jamichaelov and I will be presenting our paper at the @CogInterp workshop 13:15 - 14:45 on Dec 7th. We show how disaggregating grammatical benchmarks over the course of training reveals stages of training where models learn heuristics before learning more generalizable patterns.
New paper accepted at Findings of ACL! TL;DR: While language models generally predict sentences describing possible events to have a higher probability than impossible (animacy-violating) ones, this is not robust for generally unlikely events + is impacted by semantic relatedness
In the most extreme case, LMs assign sentences such as ‘the car was given a parking ticket by the explorer’ (unlikely but possible event) a lower probability than ‘the car was given a parking ticket by the brake’ (impossible event, related final word) over half of the time.
See the full paper here: arxiv.org/abs/2506.06808
