Unverified50% confidenceFactExact time
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
1
Sources
50%
Confidence
Long-term
Relevance
7/13/2026
First Seen
Sources
纽约时报指控OpenAI隐匿证据，ChatGPT版权诉讼再升级
rss7/9/2026
Related Claims
VerifiedResearch shows that large language models exhibit a significant "Lost in the Middle" phenomenon when processing long texts, retaining information from the beginning and end of the input text well, but showing a marked decline in attention to information in the middle.80% similarVerified大语言模型通过在数千亿乃至数万亿个词元（Token）的文本数据上进行自监督预训练79% similarUnverifiedLarge language models have two inherent weaknesses: their knowledge is cut off at training time and they cannot obtain the latest information, and they can only generate text and cannot directly perform actions.74% similarUnverifiedLarge language models can complete text rewriting operations such as synonym substitution, sentence restructuring, and tone adjustment within seconds, allowing the rewritten text to easily evade traditional text detection tools based on cosine similarity or edit distance.74% similarUnverifiedLarge language models suffer from 'attention decay' when generating long text, causing adherence to earlier instructions to decrease as context length increases73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/502987API
curl https://kongchang.com/api/v1/knowledge/claims/502987MCP
get_claim(id=502987)