GSM8K has been a cornerstone benchmark for LLMs, but performance seemed stuck around 95%. Why? Turns out, the benchmark itself was noisy. We fixed that, and found that it significantly affects evals. Introducing GSM8K-Platinum! w/Eddie Vendrow Josh Vendrow Sara Beery

Trajectory Balance with Asynchrony: Decoupling Exploration and Learning for Fast, Scalable LLM Post-Training TBA is a scalable RL system for LLM post-training that uses off-policy data and replay buffers with Trajectory Balance. It decouples training from search, improving speed

At Lawrence Livermore National Laboratory, we are using AI to: ⚛️ Solve nuclear fusion 🧪 Discover critical materials 🧠 Red-team vulnerabilities All to push science forward and protect national security 🌎 Post-training LLMs at scale can unlock these advances. But even with El Capitan—the world’s

New paper on synthetic pretraining! We show LMs can synthesize their own thoughts for more data-efficient pretraining, bootstrapping their capabilities on limited, task-agnostic data. We call this new paradigm “reasoning to learn”. arxiv.org/abs/2503.18866 Here’s how it works🧵

[LG] Trajectory Balance with Asynchrony: Decoupling Exploration and Learning for Fast, Scalable LLM Post-Training B R. Bartoldson, S Venkatraman, J Diffenderfer, M Jain... [Lawrence Livermore National Laboratory & Mila] (2025) arxiv.org/abs/2503.18929

🚨 Interested in adopting Large Reasoning Models (LRMs) but concerned about safety risks? 🚨 Meet STAR-1 🌟 – A compact, high-quality safety dataset (just 1K samples!) boosting LRMs' safety by 40% with only a minimal (~1.1%) reasoning drop! 🚀 How we built STAR-1 in just 3

🚨Concerned about visual jailbreaking attacks holding back Vision-Language Model (VLM) deployment? 🌟 Excited to announce our latest research: Double Visual Defense! TL;DR: We introduce ΔCLIP and Δ²LLaVA — the first to reconcile robust adversarial performance with

🚀 The code for LLM post-training with TBA is now available! Try out Trajectory Balance with Asynchrony via github.com/bbartoldson/TBA. x.com/bartoldson/sta…

🥳Come chat with Brian Bartoldson and Moksh Jain at our TBA poster at the #ICLR25 workshop on Open Science for Foundation Models (SCI-FM). The workshop will be held in EXPO Hall 4 #5 on Monday, April 28th.

Can you train a performant language models without using unlicensed text? We are thrilled to announce the Common Pile v0.1, an 8TB dataset of openly licensed and public domain text. We train 7B models for 1T and 2T tokens and match the performance similar models like LLaMA 1&2

Here's a free/gift link to the Washington Post article about training LLMs on openly licensed text: wapo.st/3T94IfQ. x.com/AiEleuther/sta…

🚀 Excited to introduce our latest work GRESO: a method that identifies and skips millions of uninformative prompts before rollout and achieves up to 2.0x wall-clock time speedup in training. More rollouts lead to better model performance, but they’re also a major bottleneck in