Anthropic 水印扩至更多 Claude 模型，Opus 4.6 和 Sonnet 4.6 存在绕过漏洞
Anthropic's watermark rollout leaves a loophole on two older Claude models
Anthropic 给 Claude 加了水印，但 Opus 4.6 和 Sonnet 4.6 改个采样参数就能绕过，官方只发了个合规警告。
Anthropic 为满足欧盟 AI 法案 12 月截止期限，正在把文本水印扩展到另外 4 个 Claude 模型。但在旧款 Claude Opus 4.6 和 Sonnet 4.6 上，开发者只需调整一个采样参数就能绕过水印。Anthropic 对此漏洞仅以合规警告作为回应，未做技术封堵。
Anthropic's watermark rollout leaves a loophole on two older Claude models
Anthropic is expanding its text watermark to four more Claude models to meet the EU AI Act's December deadline. On the older models Claude Opus 4.6 and Sonnet 4.6, developers can bypass the watermark by tweaking a single sampling parameter. Anthropic addresses the loophole with just a compliance warning. The article Anthropic's watermark rollout leaves a loophole on two older Claude models appeared first on The Decoder .
