Unverified50% confidenceFactExact time
NarraSpace's evaluation is based on the 48-person test set of the ADReSS benchmark, so the sample size is relatively limited, and the coverage of predefined scene events may also be constrained by the design of the image itself.
1
Sources
50%
Confidence
Long-term
Relevance
10/21/2026
First Seen
Sources
Related Entities
Related Claims
Unverified预处理泄漏在小数据集上尤为显著，当训练集仅有数百张图像时加入测试集样本会明显改变均值和标准差估计69% similarUnverifiedThe research team performed packet interventions on 18 audited masked systems and found that they exhibited intermediate-value changes on eligible cases as predicted, but the limited audit cannot establish causal mediation.65% similarUnverifiedA common limitation of red-team testing, capability evaluations, and alignment checks is that test set coverage is limited, and models may exhibit dangerous behaviors outside known test scenarios that are not captured.65% similarUnverifiedHumanEval has a limited sample size and has been extensively covered by training data, leading to suspicions of test set leakage in some models65% similarUnverifiedEvaluation sets suffer from representational bias, as they can only cover situations the builders anticipated, meaning a model that performs perfectly on the eval set may still fail on real traffic after launch.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1078833API
curl https://kongchang.com/api/v1/knowledge/claims/1078833MCP
get_claim(id=1078833)