Unverified50% confidenceFactExact time
DiscoBench数据集从影视、体育、游戏、学术、金融等11个真实领域收集多跳问答样本，最终生成211个测试样本，包含463个歧义实例
1
Sources
50%
Confidence
Long-term
Relevance
7/9/2026
First Seen
Sources
Related Claims
Unverified实验采用8:2比例划分数据，20%作为内部测试集，训练集内部使用5折交叉验证，并移除重复记录以避免数据泄漏57% similarUnverifiedWhen a team tests 1,000 combinations of strategy parameters, even if all combinations have zero predictive power, roughly 50 combinations will appear to perform excellently in backtesting purely due to randomness (the multiple comparisons problem).56% similarUnverified运行10次测试置信区间宽度约±31%，100次收窄至±10%，400次进一步收窄至±5%，推荐关键测试用例至少运行20-50次55% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/387218API
curl https://kongchang.com/api/v1/knowledge/claims/387218MCP
get_claim(id=387218)