EasyDeepLearn

How does RL contribute to jailbreak defense?

hard

Answer

  • (1) Safety-tuned RM penalizes harmful outputs.
  • (2) Adversarial RLHF: red-team probes for jailbreaks, add examples to preference data.
  • (3) Refusal tuning: SFT + RLHF on 'appropriate refusal' patterns.
  • (4) Multi-round RLAIF on constitutional harmful principles.
  • Trade-off: too aggressive → over-refusal; too lax → harmful outputs.
  • Modern practice: continuous red-teaming with automated attacks (GCG) as part of RM training loop.
Check yourself — multiple choice
  • No defense
  • Safety RM + adversarial red-team pairs + refusal tuning + constitutional RLAIF; trade-off over-refusal vs unsafe; modern: continuous red-teaming + GCG-style automated attacks in loop
  • Random
  • Only filter

Jailbreak defense: safety RM + adversarial pairs + refusal SFT + red-team GCG loop.

#llm#alignment#safety

Practise Reinforcement Learning

214 interview questions in this topic.

Related questions