Matej Balog

I'm incredibly proud to announce a productive NeurIPS for our Post-training AI Research team. We had two papers on test-time scaling accepted into the Workshop on Efficient Reasoning, contributing to a fantastic ~6 acceptances for the wider Applied AI Research @ Cerebras. Both of our papers are now available on ArXiv:  - The Conductor and the Engine: A Path Towards Co-Designed Reasoning. Link - https://lnkd.in/g5Dnnz6C. - Calibrated Reasoning: An Explanatory Verifier for Dynamic and Efficient Problem-Solving. Link - https://lnkd.in/gH2jTGeP. In the spirit of collaboration, we are also open-sourcing our updated CePO flow as part of our OptiLLM library on GitHub. This is the same methodology that achieved top scores on the Artificial Analysis Leaderboard. Link - https://lnkd.in/gd6anwFN Congratulations to the team on this incredible milestone: Anisha Garg, David Bick, Engin Tekin, Michael Wang, Pawel Filipczuk, Amaan Dhada, Yash More, Nishit N. and rest of the Post-training Team! Looking Ahead A key enabler for these results and our upcoming work in Coding Agents has been using Reinforcement Learning (RL) to provide LLMs with new, specialized expertise, which makes them highly amenable to test-time compute scaling. This brings me to our next exciting step... 🚀 Announcing Limited Early Access to our RL Service! 🚀 Is your team excited by the potential of powerful LLM models — closed source like GPT-5/Claude/Gemini or Open-source like Qwen3, GPT-OSS — but frustrated when they fail that "last mile" on your specific, critical tasks? We are opening up a limited early access program to help you solve this. Our RL service is designed to transform general models into world-class experts for your unique domain. If you're interested in building an AI system that is an expert at solving your tasks, reach out to me to see if you're a fit for the program. 📧 Email: [email protected] When you reach out, please include: - Subject: "RL Service Early Access" - Body: A brief description of your application and the "last-mile" challenges you're facing. Looking forward to post-training your problems away!

122

1 Comment

Some recent, exciting news on compressed tokenizers and how individual tokens modify images in encoder-decoder models. I hope this opens up a new avenue of research with smaller models across frameworks (perhaps JEPA?). https://lnkd.in/evFw8WBu

69

2 Comments

A bit late to the party, but happy to share our #NeurIPS2024 paper "Understanding Information Storage and Transfer in Multi-modal Large Language Models": https://lnkd.in/dq4FD5mj! We use a causal tracing methodology to study how multi-modal LLMs (MLLMs) like LLaVa and Phi retrieve information required to answer factual visual questions. Our experiments reveal that 1) MLLMs rely on earlier layers for information storage compared to LLMs, and 2) a small subset of visual tokens play a crucial role in transferring visual information to these causal blocks. We also introduce a new model-editing algorithm that leverages these findings to correct errors and insert new information into MLLMs by targeting these early causal blocks. You can find the code for our work here: https://lnkd.in/d4gr7fFB Excellent work led by Samyadeep Basu, in collaboration with Besmira Nushi, Cecily Morrison, Martin Grayson, and Soheil Feizi.

191

5 Comments