Unverified50% confidenceFactExact time
Large language models tend to assign greater attention weight to recent inputs and high-frequency words, a phenomenon known as recency bias.
1
Sources
50%
Confidence
Long-term
Relevance
9/30/2026
First Seen
Sources
Related Entities
Related Claims
Unverified大语言模型在生成摘要时天然倾向于反映评论中的多数意见，可能将低频的安全警告边缘化75% similarUnverified在大语言模型的注意力机制中，过期旧笔记和权威日志在物理上处于完全平等的竞争地位，语义更接近问题的内容可能获得更高权重73% similarUnverifiedCurrent large language models exhibit a 'Sycophancy Bias,' meaning AI tends to defer to the user's position rather than providing objectively correct answers.72% similarUnverifiedIn large language models, more input tokens cause more dispersed model attention, potentially causing key information to be overlooked72% similarUnverifiedLarge language models exhibit a sycophancy tendency, changing originally correct positions with relatively high probability when users express dissatisfaction.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/962411API
curl https://kongchang.com/api/v1/knowledge/claims/962411MCP
get_claim(id=962411)