Unverified50% confidenceFactExact time
A team from the University of Maryland proposed a lexical watermarking scheme that biases toward specific vocabulary during the token sampling process.
1
Sources
50%
Confidence
Long-term
Relevance
9/12/2026
First Seen
Sources
Related Entities
Related Claims
Unverified基于词表分割的文本水印方法由马里兰大学John Kirchenbauer等团队于2023年提出67% similarUnverifiedThe concept of AI text watermarking was first systematically proposed by John Kirchenbauer and others at the University of Maryland in their 2023 paper "A Watermark for Large Language Models".60% similarUnverifiedWatermarks in large language models are not visible messages, but a subtle statistical pattern introduced by the model when selecting tokens.56% similarUnverified口语化采样的核心思想是通过改变提示词的表述方式让大语言模型产生更多样化的输出，本质上是一种提示工程技巧54% similarUnverifiedThe training corpora of large language models may include unpublished academic preprints and research notes, raising concerns about data contamination.53% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/915947API
curl https://kongchang.com/api/v1/knowledge/claims/915947MCP
get_claim(id=915947)