Allen Institute for AI

A nonprofit AI laboratory in Seattle with a long record in language understanding and scientific search and, more recently, in fully openopen-source AI, n.: AI released, ideally with its code and details of its training data and not just the finished model, so that anyone may use, inspect, alter, share and even rebuild it. How open is open enough is debated. language models. Its projects include Semantic Scholar (2015), an AI-assisted search engine for scientific literature; the AllenNLP software library; ELMoELMo, prop. n.: A 2018 model whose word representations changed with context, so 'bank' got different numbers in 'river bank' and 'bank loan'. It was a stepping stone between fixed word vectors and BERT. (2018), whose contextual word representations preceded BERTBERT, prop. n.: An influential 2018 Google language model that learned to understand text by filling in hidden words. It made search engines and many language tools much better at grasping meaning in context.; Aristo, a system that passed eighth-grade science exams; and the OLMo language modelslarge language model, n.: A computer program trained on enormous amounts of text to predict the words that come next. At vast scale, that simple skill lets it chat, answer questions, translate, summarize and write code. (from 2024), released together with their training data (Dolma), code, training logs and intermediate checkpoints, alongside the Tülu post-training recipes and Molmo multimodal models. These fully documented releases make it a reference point for research on training data and reproducibility, in contrast to open-weightopen-weight model, n.: An AI model whose trained internal numbers are published for anyone to download and run on their own computers, even if the data and code used to make it stay private. releases that disclose only the trained parameters.