Credit: David Baillot/University of California San Diego
Toxic prompts, cloaked in benign language, can be detected far better by ToxicChat, a new benchmark developed by University of California San Diego computer scientists than by models trained on previous toxicity benchmarks.
close
-
News keyboard_arrow_right
Latest News Coronavirus News Currently Embargoed News Research News Releases Google Fact Check Research Alert Preprints Marketplace News with Video/Audio
RSS
Medicine keyboard_arrow_right
Sciencekeyboard_arrow_right
Life keyboard_arrow_right
Business keyboard_arrow_right
Journal News keyboard_arrow_right
By Location keyboard_arrow_right
Meetings, Grants, and Events keyboard_arrow_right