Multi-agent defenses reduce false negatives but increase false positives, and optimal agent count varies by jailbreak attack type.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Evaluating Multi-Agent Defences Against Jailbreaking Attacks on Large Language Models
Multi-agent defenses reduce false negatives but increase false positives, and optimal agent count varies by jailbreak attack type.