rohanpaul_ai· @rohanpaul_ai · X·· 1 天前AI 评分
AI文本检测器在新模型代际下失效
该论文发现，基于厂商旧模型训练的AI文本检测器，在模型换代前能识别超过99%的改写文本，换代后仅能识别3.8%。 用于筛查科学论文AI写作的检测器，在新一代LLM到来时可能失效，因此任何依赖它们的人都应在每次模型发布时重新测试。
This paper finds that AI-text detectors trained on a vendor's older models caught over 99% of rewrites before a generation change and only 3.8% after it.
Detectors that screen scientific papers for AI writing can stop working when a new LLM generation arrives, so anyone relying on them should re-test them with every model release.
来源：rohanpaul_ai · x.com