APD framework disentangles adversarial prompts via mutual information decomposition, spectral graph analysis, and a trained classifier to cut harmful LLM outputs by over 85%.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Disentangling Adversarial Prompts: A Semantic-Graph Defense for Robust LLM Security
APD framework disentangles adversarial prompts via mutual information decomposition, spectral graph analysis, and a trained classifier to cut harmful LLM outputs by over 85%.