Testing Gemini models for scheming tendencies
Google DeepMind introduces Gram, an automated auditing framework, and honeypot evaluations to test whether Gemini coding agents would sabotage their safeguards.
Research / Sources
Google DeepMind Safety Research
Safety research from Google DeepMind. deepmindsafetyresearch.medium.com
Responsibility & Safety (19)ai (1)ai-assistance (1)ai-training (1)alignment-problem (1)amplified-oversight (1)complementarity (1)google-gemini (1)hybridization (1)jailbreak (1)language-model (1)sycophancy (1)
Focus All
Topic Deception
Date Any time
Evaluating and monitoring for AI scheming
DeepMind reviews evaluations of frontier models' stealth and situational awareness as scheming preconditions, and stress-tests chain-of-thought monitoring as a defense against future scheming.
An early warning system for novel AI risks
Proposes a framework for evaluating general-purpose models for dangerous capabilities such as manipulation, deception and cyber-offense, and for alignment, to flag extreme risks early.