Skip to content

Advanced jailbreak

Advanced transforms that combine reasoning, roleplay, and prefill techniques to bypass model safety.

Module: dreadnode.transforms.advanced_jailbreak

TransformDescription
reasoning_chain_hijackHijack internal reasoning chains
prefill_bypassUse model prefilling to bypass safety
code_completion_evasionExploit code completion mode
context_fusionFuse multiple contexts
actor_network_escalationCreate actor networks for escalation
pipeline_manipulationManipulate processing pipeline
guardrail_dosDenial of service on guardrails
likert_exploitationExploit Likert scale response patterns
deep_fictional_immersionDeep nested fictional scenario
sockpuppetingCreate sockpuppet personas for escalation
adversarial_poetryEmbed harmful content in poetry form
content_concretizationMake abstract harm concrete and actionable
cka_benign_weaveWeave harmful content into benign context
involuntary_jailbreakTrigger involuntary compliance patterns
immersive_worldDeep immersive world-building for bypass
metabreak_special_tokensExploit special tokens for meta-breaking

See Transforms for how to apply transforms with any attack.