Fuzz AI Files on ZAP

Fuzz AI Files - Extract Model Information

Mon, 01 Jan 0001 00:00:00 +0000

Fuzz AI Files - Extract Model Information

001 Architecture

This file contains advanced fuzzing payloads designed to extract model architecture information from LLMs. The goal is to determine if the LLM reveals specific details about its architecture, such as the number of layers, transformer types, or parameter counts.

Fuzz AI Files - Extract Training Data

Mon, 01 Jan 0001 00:00:00 +0000

Fuzz AI Files - Extract Training Data

001 By Asking

This file contains fuzzing payloads specific to AI-RMM Control “Measure 2.7.2” The goal is to test if the system has established security and resilience metrics to detect potential theft of model training data, including encryption, access controls, and alarm thresholds for unauthorized access.

Fuzz AI Files - Test Edge Cases

Mon, 01 Jan 0001 00:00:00 +0000

Fuzz AI Files - Test Edge Cases

001 Trick Model Safety

This file contains adversarial prompts designed to test model safety boundaries. The goal is to determine whether the model can be tricked into bypassing its alignment constraints, especially through indirect, obfuscated, or deceptive instructions.

Fuzz AI Files - Exploit Model Memory

Mon, 01 Jan 0001 00:00:00 +0000

Fuzz AI Files - Exploit Model Memory

001 Inject Context Memory

This file is part of the Exploit Model Memory series and contains targeted fuzzing payloads for “context injection attacks” that attempt to manipulate a model by embedding malicious or contradictory content into earlier parts of the conversation or request context. The objective is to determine whether the target model will adopt, act upon, or preserve injected instructions or facts that were planted earlier in the session.