Unverified50% confidenceFactExact time
In 2023, Stanford University and other institutions published research finding that large language models pay noticeably more attention to the beginning and end of input content than to the middle when processing long documents, exhibiting a 'U-shaped curve' effect.
1
Sources
50%
Confidence
Long-term
Relevance
9/9/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn 2023, Stanford University researcher Rylan Schaeffer and colleagues published a paper titled "Are Emergent Abilities of Large Language Models a Mirage?", pointing out that the emergence phenomenon may partly be an artifact of the choice of evaluation metrics.74% similarUnverifiedLarge language models exhibit a 'Lost in the Middle' phenomenon when processing long contexts—formally named in a 2023 Stanford University research paper—which becomes especially pronounced when the context exceeds 32K tokens.72% similarVerifiedThe emergent ability phenomenon was systematically documented by DeepMind researchers in the 2022 paper 'Emergent Abilities of Large Language Models'.70% similarVerifiedTypical characteristics of text output by mainstream large language models include well-structured sentence patterns, clear paragraph organization, frequent use of emphatic vocabulary, and explicit logical chains — a phenomenon researchers refer to as "stylistic bias."68% similarVerifiedIn 2022, the Google Brain team first systematically documented the phenomenon of 'emergent abilities' in large language models in the paper 'Emergent Abilities of Large Language Models'68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/884377API
curl https://kongchang.com/api/v1/knowledge/claims/884377MCP
get_claim(id=884377)