Unverified60% confidenceBenchmarkExact time
Experimental data shows that when key information is located at the beginning and end of the input sequence, the model's retrieval accuracy can be as high as over 90%, but when located in the middle, it may drop sharply to below 60%.
2
Sources
60%
Confidence
Long-term
Relevance
7/12/2026
First Seen
Sources
Docx-CLI: An Open-Source Tool That Boosts AI Agent Word Document Efficiency by 50%
hackernewshackernews7/7/2026
Related Claims
Unverified实验显示模型对序列首尾信息的召回率比中间位置高出20%～40%78% similarUnverified在某些测试中，文档中间段信息的检索准确率比首尾段低40%以上77% similarUnverifiedOn reasoning-based question-answering tasks, the strongest closed-source models achieved accuracy exceeding 90%.77% similarUnverified参数量在1亿以下的轻量模型经领域数据微调后可达95%以上意图分类准确率，单次推理延迟控制在5-20ms以内76% similarUnverifiedTop academic deepfake detectors can see their accuracy plummet from 99% to below 60% when faced with previously unseen generative models, a phenomenon known as the cross-model generalization problem.76% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/490343API
curl https://kongchang.com/api/v1/knowledge/claims/490343MCP
get_claim(id=490343)