detectability

作者对 23 个浏览器 AI 代理进行了网站侧可检测性测试，发现只有 Manus 正确使用了 Web Bot Auth 签名标识自身，而多数代理（包括 ChatGPT 的 agent）的请求未签名且难以与人类区分，仅靠指针移动轨迹即可识别所有点击型代理。

This paper formalizes when benchmark contamination is detectable, deriving information-theoretic limits and proposing power-calibrated audits that distinguish a clean benchmark from a powerless detector. It reports two-sided empirical findings on calibration efficacy and validity gates.

This paper introduces STELA, a linguistics-aware watermarking framework for LLMs that leverages syntactic predictability via POS n-grams to balance text quality and detection robustness. The method enables publicly verifiable watermark detection without requiring access to model logits, demonstrating superior performance across typologically diverse languages (English, Chinese, Korean).