M3: multi-turn red-teaming as a finite-horizon Markov decision process

Evidence record

M3: multi-turn red-teaming as a finite-horizon Markov decision process

Read the original source ↗

Claim-level citations (2)

• supportsM3 models multi-turn jailbreak escalation as stateful black-box search: The M3 research frames multi-turn red teaming as a finite-horizon Markov decision process with an attacker model and a separate state model, and reports attack-success improvements of up to 53 percent for its tested ASR@5 comparisons.

  Abstract, Markov decision process and multi-turn ASR@5 comparison

• supportsM3 models multi-turn jailbreak escalation as stateful black-box search: The authors test an MDP-based attacker with a separate state-tracking model across two benchmarks and six target models, reporting improvements over compared multi-turn baselines of up to 53 percent in ASR@5.

  Abstract; state estimator, attacker action policy and six-model benchmark results