Unverified50% confidenceBenchmarkExact time
On Qwen-3.5-4B-Instruct, 75.6% of validity scores judged that the generated results at both poles of the principal axes accurately corresponded to their labels, with an annotator adjacent-agreement rate as high as 90.9%.
1
Sources
50%
Confidence
Long-term
Relevance
9/18/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedOn the Qwen-3.5-4B-Instruct model, the top two principal axes matched the style dimensions of human spontaneous requests with a precision of 72.8% and a macro-averaged recall of 43.6%.73% similarUnverified4Bit量化下模型的基准测试得分通常能保持原始精度的90%-95%72% similarUnverifiedQwen3.7 Max的输出质量高度依赖提示词质量，同一模型在不同提示词下的性能差异可高达30%-50%68% similarUnverifiedQwen 3 Max scored 65 points across eight tests with a score rate of 81.25%, ranking second on the reviewer's leaderboard, behind only the Claude flagship with a score rate of 82.5%.67% similarUnverifiedQwen3 27B在OpenAI HumanEval 164道Python题上取得85%通过率和99%应答率67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/931131API
curl https://kongchang.com/api/v1/knowledge/claims/931131MCP
get_claim(id=931131)