The full BBS treatment from me and @futrell.bsky.social on "How linguistics learned to stop worrying and love the LMs" is now out, with all the commentaries and our response. If you "Save PDF", it will give you the whole target article + commentary + response pdf: www.cambridge.org/core/journal...

I will miss #ACL2026 this year, but check out work from my students and collaborators! Kaijie (@kaijie-mo.bsky.social), Sebastian (@sebajoe.bsky.social), Asher (@asher-zheng.bsky.social), Lily, and Gauri will be there presenting the following: jessyli.com/acl2026

CosmicAI Leadership highlight! Explorable Universe AI Dr. @jessyjli.bsky.social is an Associate Professor, Linguistics, at UT Austin. Listen as she describes her research on AI for science and its evaluation.

Want to know *how* novel/creative your LLM response is and why? Use GENIE! 🧞

Ramya Namuduri@ramyanamuduri.bsky.social · 2mo ago

Are LLM-generated stories novel? They can have unique characters and cliché plots, or the other way around. A holistic score doesn’t help distinguish the two 😔. Meet GENIE 🧞 – a fine-grained novelty metric that tells you where and why a response is original!

Introducing Hero’s Journey, meticulously designed to test inductive generalization in a fun text game. ⚖️Verdict: all LLMs we tested trail far behind humans when induction involves generalization across procedures!

Asher Zheng@asher-zheng.bsky.social · 2mo ago

Spotting the rule from past experience is one thing; acting on it correctly is another. To find out, we introduce HERO's JOURNEY🦸‍♀️ to test for the LLMs’ inductive reasoning ability in multi-step setups. We found models show signs of rule induction, but scratch the surface.😮

New profession just dropped: pharma-morphologist #linguistics #NLP #morphology

@kaijie-mo.bsky.social · 2mo ago

“Dimicillin” isn’t real. We made it up. Yet many LLMs still call it an antibiotic. Across 9 models and 653 drugs, we find that drug-name affixes alone can drive pharmacological reasoning. Models often rely on morphology over facts. We trace this shortcut from behavior to mechanism. 🧵

New work on LLM safety in medicine 💊! Drug names have fixed morphological structure, so is that exploited by LLMs? Kaijie's paper reveals over-generalization that may pose safety risks if not handled carefully 👇

@kaijie-mo.bsky.social · 2mo ago

“Dimicillin” isn’t real. We made it up. Yet many LLMs still call it an antibiotic. Across 9 models and 653 drugs, we find that drug-name affixes alone can drive pharmacological reasoning. Models often rely on morphology over facts. We trace this shortcut from behavior to mechanism. 🧵

I'm on a new committee reviewing ARR's Responsible NLP Checklist and looking to potentially make changes. I'd love to hear others' thoughts on what is working well or needs revised, especially given it might be fresh in memory from the recent ARR cycle. 1/

We are thrilled to present a detailed report describing the system built for the AAAI-26 AI review pilot, the survey results, and a new benchmark that was created to assess the capabilities of the system. Read the full article: arxiv.org/pdf/2604.13940

New opinion piece on the interface between research on concepts and categories in minds vs. in neural network LMs! I take the position that there is much to be learned from this interface (e.g., learning about concepts from language alone) and outline some directions for future.

CosmicAI personnel contributed to the AAAI-26 AI review pilot, which generated automatic AI reviews of all research papers submitted to the conference’s main track. The AI reviews complemented human reviews. @mattlease.bsky.social @jessyjli.bsky.social @sebajoe.bsky.social Joydeep Biswas

@aaai.org · 3mo ago

We are thrilled to present a detailed report describing the system built for the AAAI-26 AI review pilot, the survey results, and a new benchmark that was created to assess the capabilities of the system. Read the full article: arxiv.org/pdf/2604.13940

QUDs going multimodal! With MQUD, we can train models to generate scientific questions❓that are inquisitive and insightful enough to be answered in the scientific paper! Huge thanks to the many paper authors who contributed to our data. Check out @yatingwu.bsky.social’s work:

Yating Wu@yatingwu.bsky.social · 3mo ago

What does a scientific figure make you wonder? 📊 We introduce MQUD: multimodal Questions Under Discussion for scientific figures. With 1,250 author-annotated questions over 245 figures from 56 papers, MQUD asks what scientific question a figure raises in context.

1k+ downloads each on the MINT empathy models since release 🔥 Encouraging to see the interest in our work! tl;dr: In multi-turn empathic dialogue, LLMs reuse the same discourse moves far more often than humans do; MINT uses RL to diversify them. Give it a try!👇 huggingface.co/hongli-zhan/...

Hongli Zhan@hongli-zhan.bsky.social · 3mo ago

New paper! 🏁 Last one from my PhD at UT Austin. LLMs sound empathic but repeat the same discourse moves turn after turn — at 2x the rate of humans. We built MINT🌿, the first RL framework for discourse move diversity in empathic dialogue. +25% empathy, −26% repetition. 📄 arxiv.org/abs/2604.11742

I had a fantastic time at the 2026 Harrington Symposium this week at UT Austin. It was wonderful to be able to dig into more science of AI with brilliant researchers across many specialties and viewpoints! Many things to think about! harrington.utexas.edu/faculty-fell...

Harrington Faculty Fellows Symposium 2026

harrington.utexas.edu

CosmicAI at our collaborative event - Harrington Faculty Fellows Symposium 2026 “Large Language Models: Advances and Applications.” Sign up for our email list to learn about our upcoming events! cosmicai.org/get-involved @jessyjli.bsky.social

In multi-turn conversation, LLMs tend to repeat the same kind of things over and over again. They could have different words, but we found them to be the *same discourse moves*! Introducing @hongli-zhan.bsky.social’s new work: novel discourse-level diversity rewards in post-training:

Hongli Zhan@hongli-zhan.bsky.social · 3mo ago

New paper! 🏁 Last one from my PhD at UT Austin. LLMs sound empathic but repeat the same discourse moves turn after turn — at 2x the rate of humans. We built MINT🌿, the first RL framework for discourse move diversity in empathic dialogue. +25% empathy, −26% repetition. 📄 arxiv.org/abs/2604.11742

Nice research! You may be interested in the small scale ($4 budget) verification performed by my personal Opus agent here: muninn.austegard.com/blog/this-tr... in which we also introduced a framing-resistant prompt to see how much that would mitigate the effetcs. 1/3

If you ask the same question with different framing/phrasing, do language models change their answers? This is super important in medicine because different info can have real consequences! Check out this new work from @hyesunyun.bsky.social

Hye Sun Yun@hyesunyun.bsky.social · 4mo ago

Patients ask LLMs medical questions — but how they phrase it matters more than it should. Our new preprint explores how different phrasings of patient health questions can lead to inconsistent conclusions, even with the same evidence. [1/6] Full Paper: arxiv.org/abs/2604.05051

Heading to #EACL2026! 🇲🇦 Friday 11a Poster Session 6: LMs struggle to perform inferences around discourse connectives aclanthology.org/2026.eacl-lo... Sunday 5p TeachingNLP workshop: new course on discourse+generation aclanthology.org/2026.teachin... w/ @kanishka.bsky.social + Daniel Brubaker

Want to know how well the models can brainstorm connections across different concepts? Super excited about @manyawadhwa.bsky.social’s work on measuring associative creativity!

Manya Wadhwa@manyawadhwa.bsky.social · 5mo ago

⚛️ Introducing CREATE, a benchmark for creative associative reasoning in LLMs. Making novel, meaningful connections is key for scientific & creative works. We objectively measure how well LLMs can do this. 🧵👇

What is the interplay between representations learned from (language) surface forms alone, and those learned from more grounded evidence (e.g.,vision)? Excited to share new work understanding “Cross-modal taxonomic generalization” in (V)LMs arxiv.org/abs/2603.07474 1/

Check out our special theme: new missions for NLP research!

EMNLP@emnlpmeeting.bsky.social · 5mo ago

📢 The First Call for Papers for EMNLP 2026 is officially out! 📝 We welcome long & short papers featuring original research on empirical methods for NLP. 🗓️ ARR Submission Deadline: May 25, 2026 🔗 Read the full CFP here: 2026.emnlp.org/calls/main_c... #EMNLP2026

Nearly 2 years ago, @jessyjli.bsky.social, @janetlauyeung.bsky.social, @valentinapy.bsky.social, and I decided that it's time to bring discourse structure to the center of NLP teaching.

Check out @asher-zheng.bsky.social's work on quantifying strategic language in dialogue, just appeared in the Dialogue and Discourse journal. We study non-cooperative moves that are subtle to capture, where modern AI still have trouble comprehending. Work w/ David_Beaver

Asher Zheng@asher-zheng.bsky.social · 6mo ago

Very excited to share that the paper w/ @jessyjli.bsky.social @DavidBeaver "Strategic Dialogue Assessment: The Crooked Path to Innocence" (used to have the name COBRA) was accepted by Dialogue and Discourse Vol 17 No.1. Check it out! 👉https://journals.uic.edu/ojs/index.php/dad/article/view/14503

“All bears have a property”, “Some bears have a property”, “Bears have a property” are different in terms of how the property is generalized to a specific bear – a great example of how language constrains thought! This holds for kids, adults, and according to our new work, (V)LMs! 🧵

🚨Be careful with LLMs when you ask health related questions -- even when the model relies on "evidence"! Kaijie's paper reveals a key weakness and the tricky balance between safety and faithfulness 👉

@kaijie-mo.bsky.social · 7mo ago

Hello world 👋 My first paper at UT Austin! We ask: what happens when medical “evidence” fed into an LLM is wrong? Should your AI stay faithful, or should it play it safe when the evidence is harmful? We show that frontier LLMs accept counterfactual medical evidence at face value.🧵

Accepted at EACL - excited about Morocco!

Kanishka Misra@kanishka.bsky.social · 10mo ago

"Although I hate leafy vegetables, I prefer daxes to blickets." Can you tell if daxes are leafy vegetables? LM's can't seem to! 📷 We investigate if LMs capture these inferences from connectives when they cannot rely on world knowledge. New paper w/ Daniel, Will, @jessyjli.bsky.social

New work to appear @ TACL! Language models (LMs) are remarkably good at generating novel well-formed sentences, leading to claims that they have mastered grammar. Yet they often assign higher probability to ungrammatical strings than to grammatical strings. How can both things be true? 🧵👇

Incredibly honored to serve as #EMNLP 2026 Program Chair along with @sunipadev.bsky.social and Hung-yi Lee, and General Chair @andre-t-martins.bsky.social. Looking forward to Budapest!! (With thanks to Lisa Chuyuan Li who took this photo in Suzhou!)

EMNLP@emnlpmeeting.bsky.social · 9mo ago

🌉 #EMNLP2026 will be October 24-29th in Budapest! 🌉 Thanks all for a great conference, and see you at the next one!

Delighted Sasha's (first year PhD!) work using mech interp to study complex syntax constructions won an Outstanding Paper Award at EMNLP! Also delighted the ACL community continues to recognize unabashedly linguistic topics like filler-gaps... and the huge potential for LMs to inform such topics!

Sasha Boguraev@sashaboguraev.bsky.social · last yr.

A key hypothesis in the history of linguistics is that different constructions share underlying structure. We take advantage of recent advances in mechanistic interpretability to test this hypothesis in Language Models. New work with @kmahowald.bsky.social and @cgpotts.bsky.social! 🧵👇!