The only way to be sure is measure leakage. Time consuming. An initial sorting with a fair degree of confidence might rely on cap value (as noted previously), & physical size vs voltage rating (paper caps tend to be larger for high voltages). But still the gray-area ones need testing to sort them.
Once you've started testing, you probably find units of same maker & shape/size that are all bad or all good & can sort those faster based on experience.
Is it worth your time? If not, maybe it's worth it to someone else.
Wayne
WB4OGM