This talk is part of the NLP Seminar Series.

Language Models: UI Design, Evaluations and Post-training

Karina Nguyen, OpenAI
Date: 11:00am - 12:00pm, May 16th 2024
Venue: Room 287, Gates Computer Science Building

Abstract

In this talk I will share some of the lessons from designing interfaces, post-training and evaluations of language models. I will also talk about some of the published work on societal impacts and the future of human-AI interaction more broadly.

Bio

Karina recently joined OpenAI as a Member of Technical Staff. She previously spent 2 years at Anthropic, working on a wide range of projects, from claude.ai UI to post-training and productionizing large as well as small models. Before she worked on R&D projects and products at the New York Times, Dropbox, and Square.