A blizzard is raging in Montreal when your friend says “Wow, the weather is amazing!” Humans easily interpret irony, while LLMs struggle at it. We propose a 𝘳𝘩𝘦𝘵𝘰𝘳𝘪𝘤𝘢𝘭-𝘴𝘵𝘳𝘢𝘵𝘦𝘨𝘺-𝘢𝘸𝘢𝘳𝘦 probabilistic framework as a solution. arxiv.org/abs/2506.09301 @ #acl2025
🧵 New paper at Findings #ACL2025@aclmeeting!
Not all documents are processed equally well. Some consistently yield poor results across many models.
But why? And can we predict that in advance?
Work with Steven Koniaev and Jackie Cheung @Mila_Quebec@McGill_NLP#NLProc
(1/n)
❗️ Confidence in text generation is tricky, as models can be confident in many, valid answers.
👀 Can we account for this without extra tuning or heuristics?
Do LLMs hallucinate randomly? Not quite. Our #ACL2025 (Main) paper shows that hallucinations under irrelevant contexts follow a systematic failure mode — revealing how LLMs generalize using abstract classes + context cues, albeit unreliably.
📎 Paper: arxiv.org/abs/2505.22630 1/n
46 Followers 67 FollowingHello! My name is Cesare (pronounced Chez-array or Chez). I'm a PhD student at McGill and Mila working on pragmatics and NLP for science.
98 Followers 149 FollowingResearch MSc @Mila_Quebec @mcgill_nlp | Research Fellow @RBCBorealis | reasoning and hallucination x evaluation and interpretability | Looking for Fall '26 PhD
780 Followers 1K FollowingScientist @wayve_ai / PhD from @mcgillu x @Mila_Quebec, advised by Doina Precup & @Yoshua_Bengio
A true friend who roasts you and learns with you
227 Followers 156 FollowingAI Scientist at TGHRI, PMCC, JDMI, UHN; Chair in AI Medical Imaging; Assistant Professor University of Toronto; Affiliate Faculty at Vector Institute
1K Followers 2K FollowingPGY4 @uofa_neurology | Research @mitcriticaldata @BIDMC_medicine | MSc @ihpmeuoft | MD @uoftmedicine | trying to fix the medical knowledge system
13K Followers 2K FollowingSVP and Head of Biomedical AI @Xaira_Thera; Associate Prof @UofT; Chief AI Officer @UHN; former PHD, CS @Stanford; opinions my own. #AI #healthcare #biology
46 Followers 67 FollowingHello! My name is Cesare (pronounced Chez-array or Chez). I'm a PhD student at McGill and Mila working on pragmatics and NLP for science.
780 Followers 1K FollowingScientist @wayve_ai / PhD from @mcgillu x @Mila_Quebec, advised by Doina Precup & @Yoshua_Bengio
A true friend who roasts you and learns with you
227 Followers 156 FollowingAI Scientist at TGHRI, PMCC, JDMI, UHN; Chair in AI Medical Imaging; Assistant Professor University of Toronto; Affiliate Faculty at Vector Institute
53K Followers 369 FollowingCo-Host, "Missing Middle". Husband. Father. Brother. Son. Economist. Housing guy. I used to do other stuff. Details in link.
6K Followers 66 FollowingLatest research in Trustworthy ML. Organizers: @JaydeepBorkar @sbmisi @hima_lakkaraju @sarahookr Sarah Tan @chhaviyadav_ @_cagarwal @m_lemanczyk @HaohanWang
49K Followers 9K FollowingI lead @Cohere_Labs. Formerly Research @Google Brain @GoogleDeepmind. ML Efficiency at scale, LLMs, ML reliability. Changing spaces where breakthroughs happen.
162K Followers 3K FollowingPersonal Account
Author: The View from Somewhere
Mastodon @[email protected]
BlueSky https://t.co/XAYRV7YPvQ
Also on LinkedIn. Less here
50K Followers 3K FollowingAI alignment + LLMs at Anthropic. On leave from NYU. Views not employers'. No relation to @s8mb. I think you should join @givingwhatwecan.
No recent Favorites. New Favorites will appear here.