Concept任一通过规则

Any-of Rule

An aggregation rule in AI evaluation that marks a result as passing if any single run passes. Its error rate decreases geometrically as p to the power of N, forming a mirror image of the unanimous-consistency rule.

Timeline (last 90 days)

Oct 13

Under the any-of rule (passing if any run passes), at p=0.10 the error rate decreases geometrically with N: 0.0010 for N=3 and 1.000e-10 for N=10.

Unverified50%

All Facts (1)

Unverified

Under the any-of rule (passing if any run passes), at p=0.10 the error rate decreases geometrically with N: 0.0010 for N=3 and 1.000e-10 for N=10.

50%

Source Articles

Three Runs Is Not a Sample Size: The Replication Fallacy in AI Evaluation and Its Computable RuleOct 13