研究揭示 Opus 5.5 等前沿模型的 AI 写作痕迹
原标题：Opus 5.5 loves to tell you ‘this matters’ (and other AI writing tells)
AI 摘要
营销公司 Graphite 的一项新研究分析了前沿大模型的写作习惯，发现尽管 em-dash、"delve" 等早期 AI 写作痕迹已被消除，各模型仍保留独特的用词与句式特征。研究以 ChatGPT 发布前的 1 万篇文章为人类对照组，让不同模型改写后对比，共识别出 1.3 万个在 AI 文本中出现频率至少为人类两倍的短语。Claude Opus 5.5 最爱说"this matters"（比人类高 116 倍），OpenAI Astra 则偏好"not simply X"式纠正性框架。Graphite 首席 AI 官 Greg Druck 表示，Claude 系列正逐渐接近人类用词分布，而 GPT 系列则在远离，且模型版本迭代只是替换痕迹而非减少痕迹。
正文节选
Now that LLM-generated prose is everywhere, human beings are eager for ways to sniff it out. While early tells like em-dashes and “delve” are long gone, researchers say there are still plenty of telltale habits that AI models fall back on when writing prose. A new study from the marketing firm Graphite looked at the writing habits of frontier models, sussing out each model’s favorite words and phrases. While old tells like em-dash use have been stamped out, models still fall back on contrast-hea