FeedSources
GD

Google DeepMind Publications

11 articles total

Go to source

Explore a selection of our recent research on some of the most complex and interesting challenges in AI.

Go to source
  •  Quantifying the Salience of Geo-Cultural Values for Pluralistic Safety Alignment
  •  Bridging the Scale Gap: Augmenting Human Red-Teaming to Uncover Latent Risks in T2I Models
  •  Going PLACES: Participatory Localized Red Teaming forText-to-Image Safety in the Global South
  •  Towards Structural Understanding of LLM Overthinking
  •  The Case for Globally Beneficial Technology
  •  Real-Time Group Dynamics with LLM Facilitation: Evidence from a Charity Allocation Task
  •  Artificial Minds, Human Disagreement: The Politics of AI Consciousness
  •  From AGI to ASI
  •  Solipsistic superintelligence is unlikely to be cooperative
  •  Realistic honeypot evaluations for scheming propensity
  •  Gram: Assessing sabotage propensities via automated alignment auditing