OLMo (language model)
≈ 2 min read
Names in other languages
No names in other languages are held for this entry yet.
Article bodies are English and are not machine-translated. A name in another language is not a translation of the article.
OLMo, short for Open Language Model, is a family of large language models developed by the Allen Institute for AI (Ai2). The first models, OLMo 7B, were released in February 20241. Alongside the model weights, Ai2 made available the training data, training and evaluation code, logs, metrics, and intermediate checkpoints used during development. The project was designed to make the development of language models more reproducible and open to scientific study. Its initial technical report contrasted OLMo with releases that provide model weights or inference software but do not disclose training data or development details. The report stated that access to these artifacts could support research into model behavior, biases, and risks.
OLMo, short for Open Language Model, is a family of large language models developed by the Allen Institute for AI (Ai2). The first models, OLMo 7B, were released in February 2024. Alongside the model weights, Ai2 made available the training data, training and evaluation code, logs, metrics, and intermediate checkpoints used during development. The project was designed to make the development of language models more reproducible and open to scientific study. Its initial technical report contrasted OLMo with releases that provide model weights or inference software but do not disclose training data or development details. The report stated that access to these artifacts could support research into model behavior, biases, and risks.
Text derived from https://en.wikipedia.org/w/index.php?oldid=1378746992 under CC BY-SA 4.0; modifications have been made.
References
- ↩Wikipedia “OLMo (language model)” release date (×1) · confidence 0.95 revision URL