Safety researcher

Mrinank Sharma

Worked on reward-model overoptimization and sycophancy, characterizing failure modes where models game the reward or flatter users.

Works in the archive (1)

“Lead” marks the first-listed author — a byline fact, not a measure of influence; co-authorship is shown as recorded.

Co-authors in the archive

Leo Gao, John Schulman, Jacob Hilton, Ethan Perez

View co-author network (5 people)

Loading graph…