Unverified50% confidenceBenchmarkExact time
Negative sample learning strategies have elevated the design success rate from the early 5–10% to over 50% in certain application scenarios
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/10/2026
First Seen
Valid until: 12/9/2026
Sources
Related Entities
Related Claims
Unverified推测解码中当草稿通过率低于50%时，实际加速效果会大打折扣甚至产生负优化74% similarUnverifiedNegative samples for contrastive learning fine-tuning should use hard negatives, such as positions in the same field but at different levels, rather than random negative samples.70% similarUnverifiedResearch shows that when the false positive rate exceeds 50%, the practical value of quality tools approaches zero.68% similarUnverifiedAnthropic's research shows that even mild alignment training can cause a 2–8 percentage point decline in model performance on certain reasoning-intensive tasks.65% similarUnverifiedAccording to public evaluations, the success rate of existing desktop agents in complex workflows is generally below 50%.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/885982API
curl https://kongchang.com/api/v1/knowledge/claims/885982MCP
get_claim(id=885982)