AI Control

Applied to Are Sparse Autoencoders a good idea for AI control? by Gerard Boxo 4d ago

Applied to Reduce AI Self-Allegiance by saying "he" instead of "I" by Knight Lee 7d ago

Applied to Measuring whether AIs can statelessly strategize to subvert security measures by Alex Mallen 11d ago

Applied to A toy evaluation of inference code tampering by Gunnar Zarncke 21d ago

Applied to The Queen’s Dilemma: A Paradox of Control by Raymond Arnold 1mo ago

Applied to Why imperfect adversarial robustness doesn't doom AI control by Raymond Arnold 1mo ago

Applied to Using Dangerous AI, But Safely? by Raymond Arnold 1mo ago

Applied to Sabotage Evaluations for Frontier Models by Buck Shlegeris 1mo ago

Applied to Win/continue/lose scenarios and execute/replace/audit protocols by Buck Shlegeris 1mo ago

Applied to Toward Safety Cases For AI Scheming by Mikita Balesni 2mo ago

Applied to Dario Amodei's "Machines of Loving Grace" sound incredibly dangerous, for Humans by Super AGI 2mo ago

Applied to A Brief Explanation of AI Control by Aaron_Scher 2mo ago

Applied to Coup probes: Catching catastrophes with probes trained off-policy by Stefan Heimersheim 2mo ago

Applied to Behavioral red-teaming is unlikely to produce clear, strong evidence that models aren't scheming by Buck Shlegeris 3mo ago

Applied to Schelling game evaluations for AI control by Olli Järviniemi 3mo ago

Applied to How to prevent collusion when using untrusted models to monitor each other by Buck Shlegeris 3mo ago

Applied to Secret Collusion: Will We Know When to Unplug AI? by schroederdewitt 3mo ago