Unverified60% confidenceBenchmarkExact time
In memory tests, retaining the full context allowed the model to accurately locate specific details mentioned by students 95% of the time, whereas after summary compression the accuracy dropped sharply to 32%.
2
Sources
60%
Confidence
Medium-term (~90 days)
Relevance
8/20/2026
First Seen
Valid until: 11/18/2026
Sources
Related Claims
UnverifiedExperimental 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%.70% similarUnverifiedAfter 15 training epochs on the demonstration dataset, the text classification model achieves an accuracy of 99%, with a total of 110,000 model parameters (16,000 for Attention, 33,000 for FFN, and the remainder for Embedding)68% similarUnverified实验采用8:2比例划分数据，20%作为内部测试集，训练集内部使用5折交叉验证，并移除重复记录以避免数据泄漏67% similarUnverifiedAn evaluation suite with 1,800 test cases has long maintained a stable aggregate pass rate of 97%, but the 64-case multilingual slice within it has a pass rate of only 61%67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/778558API
curl https://kongchang.com/api/v1/knowledge/claims/778558MCP
get_claim(id=778558)