Profile avatar
ai2.bsky.social
Breakthrough AI to solve the world's biggest problems. › Join us: http://allenai.org/careers › Get our newsletter: https://share.hsforms.com/1uJkWs5aDRHWhiky3aHooIg3ioxm
234 posts 3,457 followers 108 following
Regular Contributor
Active Commenter
Posts 11 Comments 39

How can we better understand how models make predictions and which components of a training dataset are shaping their behaviors? In April we introduced OLMoTrace, a feature that lets you trace the outputs of language models back to their full training data in real time. 🧵

submitted 18 hours ago • 1 comment

🚨 We're hiring a #ResearchScientist in #AI for Scientific Discovery at Ai2! Are you passionate about intelligent agents, data-driven discovery, and AI systems that accelerate science? Join us in shaping the future of research. 🧬🧠 Apply now: job-boards.greenhouse.io/thealleninst...

submitted 20 hours ago • 0 comments

Today we’re releasing a prototype of Genesys, an autonomous multi-agent LLM discovery system that aims to discover new types of language model architectures. We found Genesys can discover novel architectures competitive with the industry-standard transformer. 🧵

submitted 3 days ago • 1 comment

Hear Ai2 Senior Researcher @natolambert speak at @VentureBeat Transform Wednesday, 6/25! Join us at 2:30 PM PT for his talk, “A Taxonomy for Next-Generation Reasoning Models.” See you there! #VBTransform

submitted 6 days ago • 0 comments

How well can today’s models generalize, compose, or even innovate on unseen problems? OMEGA Ω is a new math benchmark that pushes LLMs beyond pattern-matching to test true mathematical reasoning. ⚡ allenai.org/blog/omega

submitted 6 days ago • 1 comment

New updates for olmOCR, our fully open toolkit for transforming documents (PDFs & images) into clean markdown. We released: 1️⃣ New benchmark for fair comparison of OCR engines and APIs 2️⃣ Improved inference that is faster and cheaper to run 3️⃣ Docker image for easy deployment

submitted 12 days ago • 2 comments

Congrats to the team! 📜🏆... Molmo and PixMo: Open Weights and Open Data for State-of-the-Art Vision-Language Models.

submitted 17 days ago • 1 comment

We are #1 on the @huggingface heatmap - this is what true openness looks like!🥇🎉 750+ models 230+ datasets And counting... Come build with us huggingface.co/spaces/cfahl...

submitted 18 days ago • 0 comments

We’ve just made it even easier to build with our open models and datasets ✨ What’s in our docs platform? ⚙️ Setup instructions 🚀 Hosting + deployment ⭐ Integration examples ⚡ Optimization tips

submitted 20 days ago • 1 comment

Congrats to Ai2's @sewonm.bsky.social won__min on the recognition! Her work is pushing the boundaries of how we can integrate private data into language model training flexibly and securely.

submitted 26 days ago • 0 comments

As we’ve been working towards training a new version of OLMo, we wanted to improve our methods for measuring the Critical Batch Size (CBS) of a training run, to unlock greater efficiency. but we found gaps between the methods in the literature and our practical needs for training OLMo. 🧵

submitted 27 days ago • 1 comment