Thrilled to share that our paper “The Non-Linear Representation Dilemma: Is Causal Abstraction Enough for Mechanistic Interpretability?” got accepted to NeurIPS 2025 as a Spotlight. 🎉

My master's thesis "Understanding the Surfacing of Capabilities in Language Models", has been awarded the ETH Medal 🏅for Outstanding Thesis. Huge thanks to my supervisors Chris Wendler Bob West! inf.ethz.ch/news-and-event… Thesis: research-collection.ethz.ch/entities/publi…

One day, someone will be bragging that they were Kanishka's first PhD student. That person could be you!

Late to the party, but very happy this paper got accepted to NeurIPS 2025 as a Spotlight! 😁 Main takeaway: Without prior assumptions about how DNNs encode concepts in their representations (eg, the linear representation hypothesis), we can claim any DNN implements any algorithm

It is PhD application season again 🍂 For those looking to do a PhD in AI, these are some useful resources 🤖: 1. Examples of statements of purpose (SOPs) for computer science PhD programs: cs-sop.org [1/4]

🚨New paper alert! 🚨 Tandem Training for Language Models arxiv.org/abs/2510.13551 Actions & thoughts of AI w/ superhuman skills will be hard for humans to follow, undermining human oversight of AI. We propose a new way to make AI produce human-understandable solutions. How?👉🧵

If you wanna do a PhD in NLP, this should be at the top of your list :)

If you are thinking a lot about CoT and multi-agent communication these days, check out Michael's work below 👇. And make sure to keep an eye on his work going forward, more great things to come! 👨🏻‍🍳

New paper: Finetuning on narrow domains leaves traces behind. By looking at the difference in activations before and after finetuning, we can interpret what it was finetuned for. And so can our interpretability agent! 🧵

I'm preparing a camera-ready for #neurips2025 and I'm confused about whether Contribution statements are allowed after page 10 🤔 I'd put it next to, e.g., the acknowledgement section. Instructions imply that they are not allowed. Does anyone know for sure? NeurIPS Conference

Working with Ethan is always amazing! If you're interested in psycholinguistics and language modelling, you should definitely apply :)