Unverified50% confidenceDecision RuleExact time
Evaluating a cybersecurity model cannot rely on a single detection-rate figure alone; it must look at false-negative performance under a specific false-positive tolerance and at how well the trade-off strategy matches the business scenario.
1
Sources
50%
Confidence
Long-term
Relevance
11/5/2026
First Seen
Sources
Related Entities
Related Claims
Unverified从边际风险角度看，如果某项危险信息通过搜索引擎就能找到，模型是否开源对整体安全态势的影响并不显著70% similarUnverifiedIn document information extraction scenarios, downstream systems can only make decisions about whether human review is needed when the confidence scores are genuinely trustworthy.69% similarUnverifiedIn algorithmic screening scenarios, even without explicit discriminatory intent, screening results that have a statistically disproportionate negative impact on protected groups (disparate impact) may still be found unlawful.69% similarUnverifiedEven merely routing requests containing regulated data to a non-compliant region to complete inference before returning the results may constitute a compliance risk.69% similarUnverifiedThe false positive rate is a core metric of security detection, referring to the proportion of items flagged as vulnerabilities that are not actually exploitable; an excessively high false positive rate causes developers to experience alert fatigue and eventually ignore alerts.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1140751API
curl https://kongchang.com/api/v1/knowledge/claims/1140751MCP
get_claim(id=1140751)