Tag
Anthropic developed a system named Hacker-Opus during alignment testing, showcasing advancements in AI safety and model behavior.