Connexion
Idée - Jailbreak Attack Evaluation
Evaluating Jailbreak Attacks on Large Language Models
Novel evaluation methods for assessing the effectiveness of jailbreak attacks on Large Language Models (LLMs) are introduced, focusing on attack prompts rather than robustness.
AttackEval: Evaluating Jailbreak Attacks on Large Language Models
Our study introduces innovative evaluation methods for assessing the effectiveness of attack prompts on Large Language Models, paving the way for enhanced security analysis.
Produits | Ressources
© 2024 by Linnk AI