The technical report introduces OLMo, an open language model developed by the Allen Institute for Artificial Intelligence. Language models have become vital in NLP research and commercial applications. However, many of these models are proprietary, with crucial details undisclosed. OLMo aims to change this by providing the research community with full access to a state-of-the-art language model, along with its training data, code, and evaluation tools. The release of OLMo is intended to empower the open research community and inspire further innovation in the field of language modeling.

 

Publication date: 1 Feb 2024
Project Page: https://github.com/allenai/OLMo
Paper: https://arxiv.org/pdf/2402.00838