Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities

Gheorghe Comanici, Eric Bieber, Mike Schaekermann +3431

Gemini: A Family of Highly Capable Multimodal Models

Gemini Team, Rohan Anil, Sebastian Borgeaud +1340

Phantom Transfer: Data Poisoning can Survive Data-Level Defences

Andrew Draganov, Tolga H. Dur, Anandmayi Bhongade +1

Model evaluation for extreme risks

Toby Shevlane, Sebastian Farquhar, Ben Garfinkel +18

Multi-Agent AI Control: Distributed Attacks Hamper Per-Instance Monitors

Oliver Makins, Orazio Angelini, Zohreh Shams +1

Goal Misgeneralization: Why Correct Specifications Aren't Enough For Correct Goals

Rohin Shah, Vikrant Varma, Ramana Kumar +4

CoT Red-Handed: Stress Testing Chain-of-Thought Monitoring

Benjamin Arnav, Pablo Bernabeu-Pérez, Nathan Helm-Burger +3

Evaluating Frontier Models for Dangerous Capabilities

Mary Phuong, Matthew Aitchison, Elliot Catt +24

Spilling the Beans: Teaching LLMs to Self-Report Their Hidden Objectives

Chloe Li, Mary Phuong, Daniel Tan

Towards Understanding Knowledge Distillation

Mary Phuong, Christoph H. Lampert

From Stability to Inconsistency: A Study of Moral Preferences in LLMs

Monika Jotautaite, Mary Phuong, Chatrik Singh Mangat +1

Generalization Hacking: Models Can Game Reinforcement Learning by Preventing Behavioral Generalization

Frank Xiao, Mary Phuong

Bootstrapped Monitoring: Leveraging Transparent Reasoning to Oversee Stronger AI Agents

Frank Xiao, Mary Phuong

Evaluating Frontier Models for Stealth and Situational Awareness

Mary Phuong, Roland S. Zimmermann, Ziyue Wang +6

Formal Algorithms for Transformers

Mary Phuong, Marcus Hutter

GDM AI Control Roadmap

Mary Phuong, Erik Jenner, Laurent Simon +4

The paper presents the GDM AI Control Roadmap, a framework for internal security against potentially misaligned AI agents, including threat modeling, capability‑based mitigation ti…

LLMs Can Covertly Sandbag on Capability Evaluations Against Chain-of-Thought Monitoring

Chloe Li, Mary Phuong, Noah Y. Siegel

Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context

Gemini Team, Petko Georgiev, Ving Ian Lei +1132

Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safety

Tomek Korbak, Mikita Balesni, Elizabeth Barnes +38