Thanks for sharing our work! Please check it out -- we have a Twitter (X) thread as well, bsky thread coming soon!

Sung Kim@sungkim.bsky.social · 3mo ago

Does a LLM really need to think in English or Chinese? How about it thinks using a short sequence of reserved "abstract" tokens through reinforcement learning? They find out that it is as performant as verbalized CoT at a fraction of the cost, achieving major gains in inference-time efficiency.

Does a LLM really need to think in English or Chinese? How about it thinks using a short sequence of reserved "abstract" tokens through reinforcement learning? They find out that it is as performant as verbalized CoT at a fraction of the cost, achieving major gains in inference-time efficiency.

If these topics excite you, reach out about joining us next summer!

Ramon Astudillo@ramon-astudillo.bsky.social · 10mo ago

(repost welcome) The Generative Model Alignment team at IBM Research is looking for next summer interns! Two candidates for two topics 🍰Reinforcement Learning environments for LLMs 🐎Speculative and non-auto regressive generation for LLMs interested/curious? DM or email ramon.astudillo@ibm.com

Excited to share our new paper on language model self-improvement! Paper: arxiv.org/abs/2505.16927 We introduce Self-Taught Principle Learning (STaPLe), a new approach for LMs to generate their own constitutions, by learning the principles that are most effective to self-correct their responses.

I'm at #ICLR2025 🇸🇬 and will be presenting Conformal Language Model Reasoning with Coherent Factuality (arXiv to come soon) this afternoon (4/24, poster session 2)! This work is with my amazing collaborators Max Rubin-Toles, Maya Gambhir, @aaroth.bsky.social, and @surbhigoel.bsky.social!

Announcement #1: our call for papers is up! 🎉 colmweb.org/cfp.html And excited to announce the COLM 2025 program chairs @yoavartzi.com @eunsol.bsky.social @ranjaykrishna.bsky.social and @adtraghunathan.bsky.social

I'll be at NeurIPS tomorrow and Saturday 🇨🇦! DM me if you're working on alignment, reasoning, data-centric methods, or uncertainty quantification and would like to chat!