Prompt Defenseis a feature in theEinstein Trust Layerthat helps minimize the risks ofjailbreakingand prompt injection attacks. These attacks occur when malicious users try to manipulate the AI model by providing unintended inputs.Prompt Defenseensures that the prompts are processed securely, protecting the system from such vulnerabilities.
* Option A(Secure Data Retrieval and Grounding) relates to ensuring that data used by AI is securely retrieved but does not address prompt security.
* Option B(Data Masking) focuses on protecting sensitive information but does not prevent injection attacks.
For more information, refer toSalesforce's Einstein Trust Layer documentationonPrompt Defenseand security features.