Expired50% confidenceBenchmarkExact time
在BrowseWebApp Bench基准上，微调后模型的准确率从22%提升至63%
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
9/7/2026
First Seen
Valid until: 9/21/2026(expired)
Sources
Related Entities
Related Claims
Unverified34B 80层模型比27B推理时间增加约18%73% similarVerified研究表明高质量的提示词可以将模型输出准确率提高30%-50%71% similarUnverified31%的精度提升数据来自GBrain方案自身的宣传口径，实际效果因数据集和场景而异69% similarUnverified在CLINC150基准测试上，80条微调数据可使向量路由达到85.9%的准确率，完整fine-tune可达96%以上68% similarUnverified移除歧义后所有模型准确率提升26.8至40.2个百分点，Gemini甚至达到81%，表明歧义是深度搜索代理最大性能瓶颈之一68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/871642API
curl https://kongchang.com/api/v1/knowledge/claims/871642MCP
get_claim(id=871642)