AIStoryHub Labs · Open Resource
AI reaches for the same words every time. This corpus tracks 758 of them, pulled from Kobak et al.'s analysis of 14 million PubMed abstracts, direct Claude and GPT testing, and thousands of chapters generated on the AIStoryHub platform.
2 of 758 entries
| Term | Confidence | Strength | |
|---|---|---|---|
| tapestry | Strong tell | 97 | |
| rich tapestry | Strong tell | 75 |
Primary study
Kobak et al. 2024: arXiv:2406.07016. Frequency analysis of 14M PubMed abstracts (2010–2024) measuring LLM-associated vocabulary inflation.
Stay updated
We'll send the JSON straight to your inbox and a short note whenever the corpus is updated or something worth reading goes up. No daily noise, unsubscribe anytime.
Want prose that avoids all 758 of these by construction, not by editing after the fact? That's what AIStoryHub is built for.