Latest research
April 2024 - OLMo 1.7–7B: A 24 point improvement on MMLU
Introducing an updated version of our 7 billion parameter Open Language Model, OLMo 1.7–7B.
April 2024 - Making a switch — Dolma moves to ODC-BY
We’re moving the Dolma dataset to the ODC-BY license. Here’s why.
March 2024 - RewardBench: The first benchmark & leaderboard for reward models used in RLHF
Introducing RewardBench, a benchmark for evaluating preference reward models.
February 2024 - Charting new waters: Skylight’s work to revolutionize maritime data annotation
How our work to create a first-of-its-kind tool helps accelerate the way AI can protect our ocean.
February 2024 - Hello OLMo: A truly open LLM
Introducing OLMo 7B, a truly open, state-of-the-art large language model.
February 2024 - OLMo: Open Language Model
Ai2 launches our first Open Language Model (OLMo), introducing our open framework for LLM experimentation.
January 2024 - The unreasonable effectiveness of easy training data
Language models can perform well on hard, domain-specific questions when trained only on easy questions.
December 2023 - AI for Earth, 2023 in review
Environmental AI advancements in 2023 that joined the fight to protect the environment.
December 2023 - Stable and fast: ACE, the Ai2 Climate Emulator
We present ACE, a 200M-parameter, autoregressive machine learning emulator of an existing global atmospheric model.