Hadas Orgad @ ICML

I'm excited to share that I'll be joining @KempnerInst @Harvard as a research fellow this September!

maybe I will live tweet the actionable interp workshop panel

Crazy amount of cool work concentrated in one room

It's really fun to walk around a poster session and literally want to stop by each one!

Hope everyone’s getting the most out of #icml25. We’re excited and ready for the Actionable Interpretability (@ActInterp) workshop this Saturday! Check out the schedule and join us to discuss how we can move interpretability toward more practical impact.

After a thunderstorm cancelled my flight, I finally made it to Vancouver for #ICML2025 and the @ActInterp workshop! DM if you want to chat about using interpretability for safer and more controllable AI. We’ll also present the Mech-Interp Benchmark (MIB) on Thu @ 11:00—come by!

Going to #icml2025? Don't miss the Actionable Interpretability Workshop (@ActInterp)! We've got an amazing lineup of speakers, panelists, and papers, all focused on leveraging insights from interpretability research to tackle practical, real-world problems ✨

Now accepted to #COLM2025! We formally define hidden knowledge in LLMs and show its existence in a controlled study. We even show that a model can know the answer yet fail to generate it in 1,000 attempts 😵 Looking forward to presenting and discussing our work in person.

VLMs perform better when answering questions about text than when answering the same questions about images - but why? and how can we fix it? We investigate this gap from a mechanistic interpretability perspective, and use our findings to close a third of it! 🧵

I'm excited that our @ActInterp workshop at @icmlconf received over 150 submissions! We had to expand our reviewer pool to accommodate all submissions. I hope this reflects a growing interest in more actionable approaches to interpretability.

A really interesting paper by Dana and others, dividing SAE features into two groups: input and output features. Output features are actually pretty useful for steering!

🚨New paper at #ACL2025 Findings! REVS: Unlearning Sensitive Information in LMs via Rank Editing in the Vocabulary Space. LMs memorize and leak sensitive data—emails, SSNs, URLs from their training. We propose a surgical method to unlearn it. 🧵👇w/@boknilev @mtutek 1/8

We received more submissions to our Actionable Interpretability workshop (@ActInterp) than expected, and we're now looking for additional reviewers! We're seeking reviewers to handle 2–3 papers between May 24 – June 7. Sign up here: forms.gle/FLToWY3keb832n… Thank you! 🙏

We got more submissions to the workshop than we anticipated, and are looking for reviewers willing to review 2-4 papers between May 24 and June 7. If you are interested, please self-nominate! Thank you 🙏 docs.google.com/forms/d/e/1FAI…

🚨 Announcing the keynote speakers in the @ActInterp workshop at #icml2025 Join us to hear these leading experts share their take on how interpretability can drive real-world impact in AI. @_beenkim @cogconfluence @byron_c_wallace @RICEric22