Unverified50% confidenceFactExact time
On the same benchmark, taking the best run versus taking the average can differ by 2.6 percentage points; Simular's data shows 0.7674 by the first clean run and 0.7938 by the best run.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/12/2026
First Seen
Valid until: 1/10/2027
Sources
Related Entities
Related Claims
Unverified在歧义点的四类行为中，'先搜再问'成功率最高，平均通过率达93.4%，远高于'直接猜'的56.5%，'搜多次再猜'为51.5%70% similarUnverifiedFor the 7B model, accuracy with the solver was 75.31% versus 72.02% for direct calculation, a paired improvement of +3.29 percentage points, with a 95% cluster interval of [-3.49, 10.38] crossing zero, making the advantage non-significant.70% similarUnverified每步60%准确率意味着连续5步后整体成功率约为7.8%（0.6的5次方）66% similarUnverifiedWhen using the multi-agent verification process, MARCH's accuracy was suppressed to 4.4%, whereas when the same model was directly asked which solution was better, accuracy reached 43.7%.65% similarUnverifiedIn the fourth round of fusion experiments, treating JEV as the 9th framework and taking the highest score yielded an overall accuracy of 81.5%, but parity dropped back to 50%.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1009482API
curl https://kongchang.com/api/v1/knowledge/claims/1009482MCP
get_claim(id=1009482)