SubTechniquediscoveryATLAS

AML.T0069.001System Instruction Keywords

What it is

Adversaries may discover keywords that have special meaning to the large language model (LLM), such as function names or object names. These can later be exploited to confuse or manipulate the LLM into misbehaving and to make calls to plugins the LLM has access to.

References

  1. https://atlas.mitre.org/techniques/AML.T0069.001

Related by meaning· 6

Nearest entities by semantic similarity across the cs-graph corpus.

ATLAS
Discover LLM System Information
ATLAS
LLM Prompt Crafting
ATLAS
LLM Prompt Obfuscation
ATLAS
LLM Data Leakage
ATLAS
LLM Trusted Output Components Manipulation
ATLAS
Delay Execution of LLM Instructions
Sourced from MITRE ATLAS — Adversarial Threat Landscape for AI Systems. Curated by Adam Lundqvist, SQUR.