📢 Working on the evaluation crisis of #LLMs? Join our #CHI2024 workshop on Human-centered Evaluation and Auditing of Language models (HEAL)! Submission is now open! CfP: heal-workshop.github.io Wesley Hanwen Deng Michelle Lam Motahhare Eslami Juho Kim Mina Lee Vera Liao #HCI #NLProc

10 days from the deadline! See you in Hawaii

How should we correct the mistakes of ML models? Can we just _tell_ them what they’re doing wrong? Paper: arxiv.org/abs/2402.03715

Before taking it out on your roommate for leaving dirty dishes out, you probably want to practice your conflict resolution skills first. Expert conflict resolution trainers, however, are EXPENSIVE. What if we practiced with an angry LLM instead? 😈 #CHI24 arxiv.org/abs/2309.12309

It's been great working with the folks from AI Vulnerability Database to launch a public version of IndieLabel, our prototype end-user auditing system (from our CSCW22 paper)! We hope this demo can seed further discussion and future work on user-driven AI audits ✨

Tired of your language model "delving" into things? Or maybe you like delving! We're working on a new interaction & method to customize language models, and we're looking for participants! If you're interested, please fill out the Google Form below. forms.gle/6hyPgvSsqRSYHp…

Chaining LLM calls can improve output quality, but navigating the massive space of task decompositions is challenging. Revisiting the established field of crowdsourcing, we distill strategies for effective LLM chain design and identify opportunities for future research. [1/11]

I'm excited to share our new paper exploring how LLM chatbots can be used to motivate physical activity behavior change! Check out our pre-print on arXiv: arxiv.org/abs/2405.06061

Frustrated with topic models? Wish emergent concepts were interpretable, steerable, and able to classify new data? Check out our #CHI2024 talk on Tues 9:45am in 316C (Politics of Datasets)! Or try LLooM, our open-sourced tool :) ✨ github.com/michelle123lam…

So grateful to work with a sterling team of organizers and to invite a group of amazing participants to our first HEAL workshop heal-workshop.github.io! Hosting academic parties always gives me the greatest sense of community :-) Excited to bring HEAL to future HCI/NLP venues✨

How can end users more powerfully shape LLM behavior? DITTO lets a user provide just a handful of demonstrations to align a language model to their needs—and users much prefer these results over those of baseline methods and self-authored prompts!

Our good friend Farnaz Jahanbakhsh (@[email protected]) is going off to Computer Science and Engineering at Michigan as an Assistant Professor! We are so lucky to have been able to share time with Farnaz-- she is a brilliant and talented researcher, a fun and empathetic friend, and overall great person to be around ❤️ We will miss her dearly!