Unverified50% confidenceBenchmarkExact time
The framework achieves 68% accuracy on QUBOBench, surpassing the direct single-call baseline method by 22 percentage points.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/12/2026
First Seen
Valid until: 12/11/2026
Sources
Related Entities
Related Claims
UnverifiedNVFP4量化方案相比BF16基线实现了1.5倍的推理加速，同时在top-1准确率上保留了92%到97%的水平69% similarUnverified该量化模型相较于BF16全精度版本保留了约77%的准确率68% similarUnverifiedThe fine-tuned Qwen3.5-4B model improved accuracy on BU Bench V1 from 15% to 53%.68% similarUnverifiedPrecision Mode leads the strongest frontier model chunk-and-merge baseline solution by 7 percentage points in accuracy.68% similarUnverifiedBy fine-tuning Qwen3.5-4B on approximately 3,000 browser operation trajectories, a developer improved accuracy on the BrowseWebApp Bench benchmark from 22% to 63%.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/916297API
curl https://kongchang.com/api/v1/knowledge/claims/916297MCP
get_claim(id=916297)