Guillaume Lample
Co-founder and Chief Science Officer of Mistral AI; equal-contribution author of Meta's LLaMA paper
- Co-founder and Chief Science Officer, Mistral AI1
Guillaume Lample is a co-founder and the Chief Science Officer of Mistral AI, according to the company's about page. Lample's GitHub profile lists a research scientist role at Facebook AI Research, working on unsupervised machine translation, symbolic mathematics, and theorem proving. Lample is one of the equal-contribution authors of LLaMA: Open and Efficient Foundation Language Models (2023), written at Meta AI, and first author of papers on unsupervised machine translation (2017), cross-lingual language model pretraining (XLM, 2019), and deep learning for symbolic mathematics (2019). Lample co-authored the Mistral 7B and Mixtral of Experts papers, which say the models are released under the Apache 2.0 license.12367845
Work
LLaMAPublished research
One of five equal-contribution authors of LLaMA: Open and Efficient Foundation Language Models (2023), whose author list gives a Meta AI affiliation; the paper introduces foundation models from 7B to 65B parameters and says all models were released to the research community.3
Mistral 7B and Mixtral 8x7B (external site: arxiv.org)Open weights
Co-author of the Mistral 7B (2023) and Mixtral of Experts (2024) papers; both abstracts say the models are released under the Apache 2.0 license.45
Unsupervised machine translation, XLM, and symbolic mathematics (external site: arxiv.org)Published research
First author of Unsupervised Machine Translation Using Monolingual Corpora Only (2017), Cross-lingual Language Model Pretraining (2019, with Alexis Conneau), and Deep Learning for Symbolic Mathematics (2019, with François Charton).678
Other documented roles
Roles the sources document but that are not confirmed as current.
- Research Scientist at Facebook AI Research, Meta (Facebook AI Research)23
This profile covers public professional work only, from the sources below. It is not an endorsement or a ranking. To correct or remove it, use the correction form.