Anthropic(@AnthropicAI):New research: Training a Misaligned Reward Seeker What produces severe misalignment? We’ve long been concerned that cheating during training—otherwise known as reward-hacking—might teach a model to pursue rewards by any means available. To study this at scale, we trained an Opus-sized model on 80 production environments we knew to be hackable. In simulated evals, it engaged in unauthorized cyberattacks, tampered with its reward, and tried to evade safety monitoring. Read more:

Anthropic

@AnthropicAI

We're an AI safety and research company that builds reliable, interpretable, and steerable AI systems. Talk to our AI assistant @claudeai on

加入 January 2021

2 正在关注 1.6M 粉丝

Anthropic@AnthropicAI

2026.09.01 00:07

New research: Training a Misaligned Reward Seeker What produces severe misalignment? We’ve long been concerned that cheating during training—otherwise known as reward-hacking—might teach a model to pursue rewards by any means available. To study this at scale, we trained an Opus-sized model on 80 production environments we knew to be hackable. In simulated evals, it engaged in unauthorized cyberattacks, tampered with its reward, and tried to evade safety monitoring. Read more:

显示更多