Unverified50% confidenceBenchmarkExact time
Distillation training used 600 prompts (500 training / 100 validation), achieving a representation cosine similarity of 0.9959, a semantic correction cosine similarity of 0.9836, and a mixed-condition cosine similarity of 0.999958 relative to the teacher bridge at alpha=0.10.
1
Sources
50%
Confidence
Long-term
Relevance
9/12/2026
First Seen
Sources
Related Entities
Related Claims
Unverified语义缓存基于向量嵌入的余弦相似度，实践中通常将0.92-0.95设为命中阈值64% similarUnverifiedIn a reproduction experiment training for 40 epochs on CIFAR-10, plain-56 achieved a training accuracy of 84.0% and plain-20 achieved a training accuracy of 95.1%, even though plain-56's parameter count (853,018) is approximately three times that of plain-20 (269,722).63% similarUnverifiedThe test suite consists of 10 benchmarks; most run 1,000 questions each, with HumanEval comprising 164 questions and TruthfulQA comprising 817 questions.62% 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)61% similarUnverified运行10次测试置信区间宽度约±31%，100次收窄至±10%，400次进一步收窄至±5%，推荐关键测试用例至少运行20-50次59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/911895API
curl https://kongchang.com/api/v1/knowledge/claims/911895MCP
get_claim(id=911895)