Unverified50% confidenceFactExact time
Highly repeated sentences tend to be famous quotes repeatedly cited on the internet, and the two corpora are highly consistent in which sentences are highly repeated, making it impossible to distinguish training exposure from fame.
1
Sources
50%
Confidence
Long-term
Relevance
9/11/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn word-substitution experiments, regardless of whether the original sentence appeared once or a hundred times in the corpus, the model's preference gap for the original sentence was the same, indicating that the author's phrasing rather than memorization is being rewarded.68% similarUnverifiedApproximately 15–25% of the content in spontaneous speech consists of filler words, sentence repetitions, self-corrections, and other redundancies.68% similarUnverified当训练数据中某些文本出现频率极高时，大语言模型可能产生记忆化（Memorization）现象，能够近乎逐字输出特定段落65% similarUnverified文本对抗扰动（Adversarial Text Perturbation）源自深度学习领域的对抗样本研究，通过同义词替换、句式重构等打破AI文本统计规律63% similarUnverified研究者发现ChatGPT倾向于过度使用delve、furthermore、notably等学术词汇以及It's important to note that等句式63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/907914API
curl https://kongchang.com/api/v1/knowledge/claims/907914MCP
get_claim(id=907914)