Vienna, Austria – Cohere and Cohere For AI, led by Sara Hooker, are set to have a significant presence at the 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025), taking place from July 27 to August 1, 2025. The technical team from Cohere and Cohere_Labs will be actively participating, with a focus on advancing research in large language models and multilingual AI.
A key highlight of their involvement includes the presentation of the paper "M-RewardBench: Evaluating Reward Models in Multilingual Settings," co-authored by Sara Hooker and other researchers. This research aims to enhance the evaluation of reward models in diverse linguistic environments, contributing to more robust and equitable AI systems. The paper underscores Cohere For AI's commitment to addressing critical challenges in the field.
Sara Hooker, Head of Cohere For AI, has consistently emphasized the organization's mission to make artificial intelligence accessible across a broader spectrum of languages. This initiative seeks to bridge the gap in AI representation for under-resourced languages, fostering global inclusivity in AI development. The work presented at ACL 2025 aligns directly with these strategic objectives.
The presence of Cohere's technical team at such a prominent academic conference signifies the company's dedication to open research and collaboration within the NLP community. Their participation allows for the sharing of cutting-edge developments and engagement with leading experts from around the world. As Sara Hooker stated in a recent tweet, > "Come find us at @aclmeeting 🔥✨ Plenty of the technical team from @cohere @Cohere_Labs around."
Cohere, a Canadian multinational technology company, specializes in large language models and AI products tailored for enterprise solutions across various regulated industries. Their engagement at ACL 2025 further solidifies their role as a significant contributor to both fundamental AI research and its practical applications. The insights shared at the conference are expected to influence future directions in multilingual natural language processing.