10:20Deceptive Misaligned Mesa-Optimisers? It's More Likely Than You Think...Robert Miles AI Safety93k5 yr ago
23:24The OTHER AI Alignment Problem: Mesa-Optimizers and Inner AlignmentRobert Miles AI Safety276k5 yr ago
17:52Training AI Without Writing A Reward Function, with Reward ModellingRobert Miles AI Safety259k6 yr ago
11:32How to Keep Improving When You're Better Than Any Teacher - Iterated Distillation and AmplificationRobert Miles AI Safety181k7 yr ago
5:20AI learns to Create ̵K̵Z̵F̵ ̵V̵i̵d̵e̵o̵s̵ Cat Pictures: Papers in Two Minutes #1Robert Miles AI Safety49k9 yr ago
9:38What Can We Do About Reward Hacking?: Concrete Problems in AI Safety Part 4Robert Miles AI Safety118k9 yr ago
7:32Reward Hacking Reloaded: Concrete Problems in AI Safety Part 3.5Robert Miles AI Safety96k9 yr ago
5:51The other "Killer Robot Arms Race" Elon Musk should worry aboutRobert Miles AI Safety101k9 yr ago
3:23Avoiding Positive Side Effects: Concrete Problems in AI Safety part 1.5Robert Miles AI Safety55k9 yr ago
9:33Avoiding Negative Side Effects: Concrete Problems in AI Safety part 1Robert Miles AI Safety160k9 yr ago
