A single safety demonstration appended at inference time mitigates many-shot jailbreak attacks by counteracting implicit malicious fine-tuning on harmful examples.
Knowledge-driven multi-turn jailbreaking on large language models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Mitigating Many-shot Jailbreak Attacks with One Single Demonstration
A single safety demonstration appended at inference time mitigates many-shot jailbreak attacks by counteracting implicit malicious fine-tuning on harmful examples.