works on
large language models 1multi-turn jailbreak 1policy optimization 1reinforcement learning 1turn-level credit assignment 1

From the 1 of 2 linked papers with an AI index.

collaborators
Showing cs.CLShow all

1 paper · 1 filter

cs.CL2026

MJ: Multi-turn LLM Jailbreaking via Decomposed Credit Assignment

Junyoung Park, Namgyu Park, Sechan Lee +3

The paper proposes a turn-level credit assignment framework (DC‑GRPO) for training multi‑turn jailbreak attacks on large language models, showing higher success rates than prior me…