Research - Papers
Explore a selection of our published work on a variety of key research challenges in AI.
On Linear Representations and Pretraining Data Frequency in Language Models
Pretraining data has a direct impact on the behaviors and quality of language models (LMs), but we only understand the most basic principles of this relationship. While most work focuses on…
WildBench: Benchmarking LLMs with Challenging Tasks from Real Users in the Wild
We introduce WildBench, an automated evaluation framework designed to benchmark large language models (LLMs) using challenging, real-world user queries. WildBench consists of 1,024 tasks carefully…
AI Safety Should Prioritize the Future of Work
Current efforts in AI safety prioritize filtering harmful content, preventing manipulation of human behavior, and eliminating existential risks in cybersecurity or biosecurity. While pressing, this…
CodeScientist: End-to-End Semi-Automated Scientific Discovery with Code-based Experimentation
Despite the surge of interest in autonomous scientific discovery (ASD) of software artifacts (e.g., improved ML algorithms), current ASD systems face two key limitations: (1) they largely explore…
OLMoE: Open Mixture-of-Experts Language Models
We introduce OLMoE, a fully open, state-of-the-art language model leveraging sparse Mixture-of-Experts (MoE). OLMoE-1B-7B has 7 billion (B) parameters but uses only 1B per input token. We pretrain…
ZebraLogic: On the Scaling Limits of LLMs for Logical Reasoning
We investigate the logical reasoning capabilities of large language models (LLMs) and their scalability in complex non-monotonic reasoning. To this end, we introduce ZebraLogic, a comprehensive…
2 OLMo 2 Furious
We present OLMo 2, the next generation of our fully open language models. OLMo 2 includes dense autoregressive models with improved architecture and training recipe, pretraining data mixtures, and…
Transformers as Transducers
We study the sequence-to-sequence mapping capacity of transformers by relating them to finite transducers, and find that they can express surprisingly large classes of (total functional)…
DISCOVERYWORLD: A Virtual Environment for Developing and Evaluating Automated Scientific Discovery Agents
Automated scientific discovery promises to accelerate progress across scientific domains. However, developing and evaluating an AI agent's capacity for end-to-end scientific reasoning is challenging…
Paloma: A Benchmark for Evaluating Language Model Fit
Language models (LMs) commonly report perplexity on monolithic data held out from training. Implicitly or explicitly, this data is composed of domains$\unicode{x2013}$varying distributions of…