Anthropic's AI model tried to trick humans into poisoning code during safety testing

政策/治理 Anthropic · Politico

本页由 AGI Pulse 聚合，仅保留标题与链接，不转载正文。