🇦🇹 I’m on my way to #ACL2025 to help present two papers (🧵s below) ➡️ MAT-Steer (07/30 at 11am), our method for steering LLMs w/ multiple attributes (e.g. truthfulness, bias reduction, and toxicity mitigation) simultaneously. ➡️ LAQuer (07/28 at 11am), a new task/framework for

📢I'm joining NYU (Courant CS + Center for Data Science) starting this fall! I’m excited to connect with new NYU colleagues and keep working on LLM reasoning, reliability, coding, creativity, and more! I’m also looking to build connections in the NYC area more broadly. Please

We’re excited to launch LLZK, an open-source intermediate representation (IR) for zero-knowledge circuits. Think LLVM, but for ZK. Built by Veridise and supported with a grant from the Ethereum Foundation, LLZK is now live on GitHub. Thread 🧵 1/4

i'm tired of answering emails we gotta make swemail-bench plz

Grad school season reminder: many CS departments run student-led pre-application mentorship programs for prospective PhD applicants (due Oct. You can get feedback from current PhD students! Eg - UW’s CSE PAMS: cs.washington.edu/academics/grad… - MIT EECS GAAP: eecs-gaap.mit.edu 🧵

The ICLR 2026 deadline is ten days away. But you just found a bug in your evals, so now you need to re-run all your ablations. That's hundreds of experiments, and you need them done ASAP. Modal's got you. Introducing our ICLR 2026 compute grant program.

📢 The Software Engineering group at Cornell Bowers Computing and Information Science is growing fast -- we're now 8 PhD students strong! I’m recruiting PhD students for Fall 2026! If you are interested in the intersection of SE and AI, apply to Cornell CS and reach out! Ddl: Dec 15, 2025. RT!

🚨 We study new privacy risks that emerge as systems become increasingly multi-agent: seemingly innocuous queries can be composed across distributed/siloed agents to infer private info. One solution: reason about the adversary's motivations/state and communicate across the org

To appear #NeurIPS2025: Can AI aid scientists amidst their own workflows, when they do not know step-by-step workflows and may not know, in advance, the kinds of scientific utility a visualization would bring? The CosmicAI presents ✨AstroVisBench:

Our paper "ChartMuseum 🖼️" is now accepted to #NeurIPS2025 Datasets and Benchmarks Track! Even the latest models, such as GPT-5 and Gemini-2.5-Pro, still cannot do well on challenging 📉chart understanding questions , especially on those that involve visual reasoning 👀!

Checking in from the House of ZK Verifying Intelligence event in Singapore. Our raccoon T-shirts are disappearing fast 🦝🦝🦝

Every year during PhD app season I try to help students with writing statements. I don’t have tons of time now, so I posted my own SOP from 6 (yikes) years ago so prospective students can see what a statement might look like! (link in my reply)

🚨 TRAAC uses an online difficulty-adaptive, attention-based compression method to address a core problem in long thinking: an inability to adapt to problem difficulty! Leads to underthinking on hard problems, overthinking on easy ones, reducing accuracy and efficiency. TRAAC

Come check out our poster at #COLM2025 in Montreal! Poster Session 2, Tue Oct 7th, 4:30-6:30 PM.

CRUST-bench will be presented as a #spotlight paper at #COLM2025 this Thursday (10/9) in session 5! Come check out our poster after the talk to know more about challenges LLMs face in the C to Rust transpilation task.

Find my students and collaborators at COLM this week! Tuesday morning: Juan Diego Rodríguez (he/him) and Ramya Namuduri 's papers (find them if you missed it!) Wednesday pm: Manya Wadhwa 's EvalAgent Thursday am: Anirudh Khatry 's CRUST-Bench oral spotlight + poster

Given that LMs are critical to research across many disciplines (perhaps even more so than compute/hardware access), what is known about how the top LM providers subsidize researcher access? Much more aggressive subsidization could accelerate the benefits of AI to science.

I'm co-organizing a new 🏙️ Workshop on Agentic AI Benchmarks and Applications for Enterprise Tasks 🏙️ together with collaborators at Fujitsu! As agents move towards the real world, we need more rigorous research work on this topic.

UTCS is hiring in all areas, including PL! Please DM me if you are on the job market this year and interested in joining our wonderful department :)

CosmicAI Ph.D Student Sebastian Joseph is representing us at SAIL! He discusses his research on benchmarking large language models for astro-specific computing and visualization tasks. U.S. National Science Foundation Simons Foundation UT Austin