CoinWorld News: Blockchain security firm OpenZeppelin's audit found that the AI security benchmark EVMbench, launched through a collaboration between OpenAI and Paradigm, has training data contamination and misclassification of vulnerabilities. At least four vulnerabilities marked as high severity are actually not exploitable, which may affect the accuracy of the assessment.
View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
CoinWorld News: Blockchain security firm OpenZeppelin's audit found that the AI security benchmark EVMbench, launched through a collaboration between OpenAI and Paradigm, has training data contamination and misclassification of vulnerabilities. At least four vulnerabilities marked as high severity are actually not exploitable, which may affect the accuracy of the assessment.