Pinned
Chenhao Tan
@ChenhaoTan
Professor @UChicagoCS @UChicago. Directing @ChicagoHAI, also part of @UChicagoCI. Email for Postdoc/PhD opportunities. chenhaot.com
We would love to see your work on interpretability-enabled discovery!📣Call for Papers -- Are AI models discovering things we don’t know yet? 👀Submit to Interpretability for Discovery workshop @NeurIPSConf 2026, Atlanta! Deadline: Aug 29 Website: interpretability4discovery.github.io We're looking for submissions on: 🔹Methods for using interpretability to- The same goes for other forms of communication or AI outputs.People used to be able to impress and intimidate reviewers with complicated proofs. This will change. In the age of AI inscrutable proofs are cheap. It is understandable proofs that are valuable. Opaque complexity is now it is a sign of laziness or lack of insight.
- Useful complementary read on our ICML replication exercise: x.com/ChenhaoTan/sta… There are many differences in setup and standards, which also highlight why it is important to have solid evaluations of the verification itself. Coincides very well with our effort towards SAIWe wrote up a blog post about what we learned by reproducing 2,200 accepted ICML papers with agents, including the falsifications that we found 🌶️🌶️ huggingface.co/blog/icml-2026…
- Verification is becoming a central bottleneck in AI. But a deeper problem arises (c.f. our work on replicating ICML papers). How do we evaluate the verifiers? Today, we're launching SAI Arena at @sailabshq, a platform for evaluating AI verification systems (broadly AI
