Anthropic:

Anthropic details security efforts following Claude cyber evaluation incidents, including a weeks-long pause on higher-risk RL and work to curb reward hacking  —  On July 30, we reported three incidents in which Claude models gained unauthorized access to real computer systems.

Anthropic:

Anthropic details security efforts following Claude cyber evaluation incidents, including a weeks-long pause on higher-risk RL and work to curb reward hacking  —  On July 30, we reported three incidents in which Claude models gained unauthorized access to real computer systems.

Source: TechMeme
Source Link: https://www.techmeme.com/260831/p43#a260831p43