🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Hi friends, colleagues, followers. I am on the faculty job market! I am a PhD student Berkeley School of Information + Berkeley AI Research. I work on NLP, and I believe all language, whether AI- or human-generated, is ✨social and cultural data✨. My work includes: 🧵

I'm recruiting PhD students + postdocs for my lab, coming to Johns Hopkins University in Fall 2025! Our brand new lab is at the intersection of cognitive science and AI, using computational + behavioral methods to understand how language works in minds and machines. Details below! (1/4)

tinlab at Boston University (with a new logo! 🪄) is recruiting PhD students for F25 and/or a postdoc! Our interests include meaning, generalization, evaluation design, and the nature of computation/representation underlying language and cognition, in both humans and machines. ⬇️

I'll be presenting this paper next week at EMNLP. If you are interested in reward model generalizability and/or multilingual/cross-lingual alignment (or any random stuff), I'd be happy to chat!

💡We find that models “think” 💭 in English (or in general, their dominant language) when processing distinct non-English or even non-language data types 🤯 like texts in other languages, arithmetic expressions, code, visual inputs, & audio inputs ‼️ 🧵⬇️arxiv.org/abs/2411.04986

Why do we treat train and test times so differently? Why is one “training” and the other “in-context learning”? Just take a few gradients during test-time — a simple way to increase test time compute — and get a SoTA in ARC public validation set 61%=avg. human score! ARC Prize

If you like➕🔢💭 or you like 🐟, visit Kyle Lo Wed morning (todayyy) at #EMNLP2024 🌴. There will be a poster on this paper! (real 🐟 not included)

Everyone talks about AI tutors replacing humans...But can AI even follow a real tutoring session? Introducing POSR: The first framework & dataset spanning over 24,300 mins of real human tutoring to evaluate if AI can truly understand structure in conversations! 🎯

Very cool new paper + dataset from Rose et al.!

Reminder that I'm recruiting a couple of PhD students! The deadline is Dec 6 (data science) or Dec 18 (linguistics). Keywords: cognitively plausible LMs, generalization from small data, linguistics-adjacent LM evaluation and analysis, human sentence processing...

I’m hiring PhD students in computer science at Columbia! Our lab will tackle core challenges in understanding and controlling neural models that interact with language. for example, - methods for LLM control - discoveries of LLM properties - pretraining for understanding

This year, I have an exceptional student on the academic market. Wenting Zhao (Wenting Zhao) builds systems that reason in natural settings. She combines AI & NLP to study newly emerging problems. She recently released WildChat (wildchat.allen.ai) and Commit-0

My superstar labmates are on the job market! 🤩 Leshem (Legend) Choshen 🤖🤗 Pratyusha Sharma Ekin Akyürek

[1/x] can we scale small, open LMs to o1 level? Using classical probabilistic inference methods, YES! Joint MIT CSAIL / Red Hat AI Innovation Team work introduces a particle filtering approach to scaling inference w/o any training! check out …abilistic-inference-scaling.github.io

Check out Isha Puri's cool new work on inference-time scaling!

MIT NLP is now on twitter! 👋

Self-Steering Language Models

Tackling complex problems with LMs requires search/planning, but how should test-time compute be structured? Introducing Self-Steering, a new meta-reasoning framework where LMs coordinate their own inference procedures by writing code!

Exciting news: the Fairness, Accountability, Transparency and Ethics (FATE) group at Microsoft Research NYC is hiring a predoctoral fellow!!! 🎉 microsoft.com/en-us/research…

here are three awesome researchers everyone should follow: - Songlin (Songlin Yang / phd at MIT), - Will William Merrill (phd at NYU / gonna be a prof soon) - Rulin (Rulin Shao / phd at UW)! and here is why: 1. if i had to bet on one person to develop an architecture that