🔥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

I’m in Singapore for #EMNLP2023! Happy to chat with folks about explainability in IR, mechanistic interpretability, user studies/human evaluation, and food! 🇸🇬

Very fortunate to be able to attend a conference with so many lab mates who are all doing super cool work!

Our #ICLR2024 paper was accepted as a spotlight: We look at whether language models reuse attention heads for functionally similar processes across different tasks. Basically, whether LMs implement reusable ‘functions’ in their weights

Our lab is looking for a passionate postdoctoral researcher with a background in language modeling. Come join us in Tübingen! uni-tuebingen.de/en/university/… #hiring #nlproc #nlp #ai #academicjobs Universität Tübingen

Calling all academic AI researchers! 🚨 We are conducting a survey on compute resources. We want to help the community better understand our capabilities+needs. We hope that this will help us all advocate for the resources we need! Please contribute at: forms.gle/3hEie4hj999fiS…

Mechanistic interpretability has advanced our understanding of LLMs, but what about multimodal models? Introducing NOTICE. NOTICE is a mech interp pipeline for multimodal models that performs activation patching without relying on Gaussian noise for corruption [1/5]. Michal Golovanevsky

I’m presenting our explainability papers today!👇First stop by at 10:30 during T1.1 (Senate) to learn about our user-centric explainability metric, and then at 1:30 during T2.3 (South American B) for our perspective on new directions for XIR research!

That’s a wrap on #sigir2024! A little (read: very) nerve wracking to be presenting for the first time, but very grateful for all the interesting discussions that followed! Now all that’s left is to see if I’ll ever make it back to PVD…

At 15:00 on 30 Sept, Catherine Chen Catherine Chen from Brown University will give a talk entitled "Advancing Explainable Information Retrieval: Methods for Human-Centered Evaluation and Interpreting Neural IR Models". Details at: samoa.dcs.gla.ac.uk/events/viewtal… UofG Computing Science Glasgow IR Group

I'm recruiting PhD students for our new lab, coming to Boston University in Fall 2025! Our lab aims to understand, improve, and precisely control how language is learned and used in natural language systems (such as language models). Details below!

Wondering how long it takes to train a 1B-param LM from scratch on your GPUs? 🧵 See our paper to learn about the current state of academic compute and how to efficiently train models! Use our code to test your own models/GPUs! arxiv.org/abs/2410.23261 github.com/apoorvkh/acade…

If SOTA models fail to recognize simple shapes, should we be evaluating them on complex geometric tasks? Most MLLMs struggle with counting the number of sides of regular polygons and all MLLMs receive 0% on novel shapes. William Rudman Amir Bar Vedant Palit [1/6]

Last call for extended abstract submissions to our workshop on explainable IR! Join us in Italy this summer for some interesting discussions and tasty food :)

I’m in Italy for SIGIR! 🇮🇹 If you’re interested in explainability/interpretability, come by our poster this afternoon (Towards Best Practices of Axiomatic Activation Patching in IR), or our workshop on Explainability in IR on Thursday, would love to chat!