Search papers, labs, and topics across Lattice.
The paper introduces Safety Self-Play (SSP), a reinforcement learning framework where a single LLM acts as both attacker and defender to improve safety alignment. SSP uses an experience replay mechanism with UCB sampling to prioritize learning from failure cases. Experiments show that SSP significantly outperforms models trained on static adversarial datasets, demonstrating its ability to evolve robust defense capabilities.
LLMs can be made significantly more robust to jailbreaking by having them red-team themselves via self-play, dynamically evolving attack strategies to uncover vulnerabilities.
Large Language Models (LLMs) have achieved remarkable capabilities but remain vulnerable to adversarial ``jailbreak''attacks designed to bypass safety guardrails. Current safety alignment methods depend heavily on static external red teaming, utilizing fixed defense prompts or pre-collected adversarial datasets. This leads to a rigid defense that overfits known patterns and fails to generalize to novel, sophisticated threats. To address this critical limitation, we propose empowering the model to be its own red teamer, capable of achieving autonomous and evolving adversarial attacks. Specifically, we introduce Safety Self- Play (SSP), a system that utilizes a single LLM to act concurrently as both the Attacker (generating jailbreaks) and the Defender (refusing harmful requests) within a unified Reinforcement Learning (RL) loop, dynamically evolving attack strategies to uncover vulnerabilities while simultaneously strengthening defense mechanisms. To ensure the Defender effectively addresses critical safety issues during the self-play, we introduce an advanced Reflective Experience Replay Mechanism, which uses an experience pool accumulated throughout the process. The mechanism employs a Upper Confidence Bound (UCB) sampling strategy to focus on failure cases with low rewards, helping the model learn from past hard mistakes while balancing exploration and exploitation. Extensive experiments demonstrate that our SSP approach autonomously evolves robust defense capabilities, significantly outperforming baselines trained on static adversarial datasets and establishing a new benchmark for proactive safety alignment.
