Language Models (Mostly) Know What They Knowwith Saurav Kadavath, Tom Conerly, Amanda Askell, Tom Henighan +312 versionsW_579nsf6fv2 · current
Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons LearnedW_rp4q6r33v2
WOMBAT v2.S: A Bayesian inversion framework for attributing global CO2_22 flux components from multiprocess dataJosh Jacobson, Michael Bertolacci, Andrew Zammit-Mangion et al.0 citations2503.09065v12025arXiv
Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons LearnedDeep Ganguli, Liane Lovitt, Jackson Kernion et al.127 citations2209.07858v22022arXivon Valency
Language Models (Mostly) Know What They KnowSaurav Kadavath, Tom Conerly, Amanda Askell et al.187 citations2207.05221v42022arXivon Valency