Oyster-II: Reinforcement Learning for Constructive Safety Alignment in Large Language Models

Researchers propose Oyster-II, a reinforcement learning framework for constructive safety alignment in large language models, addressing limitations of refusal-based strategies and supervised fine-tuning. The framework outperforms Qwen3-14B and Oyster-I on safety benchmarks, achieving cross-scale performance comparable to larger models.

arXiv:2607.02914v1 Announce Type: new Abstract: Large language models LLMs have demonstrated remarkable capabilities across diverse applications, yet ensuring their simultaneous safety, helpfulness, and trustworthiness remains a persistent challenge. Conventional refusal-oriented alignment strategies mitigate harmful content generation but systematically fail to serve legitimate user needs, often withholding information that could safely and constructively address the underlying intent of sensitive queries. Building upon the constructive safety paradigm pioneered by Oyster-I, which moves beyond blanket refusal toward thoughtful, response-oriented safety alignment, we identify two critical limitations of its Supervised Fine-Tuning SFT -based scheme: insufficient safety generalization to out-of-distribution scenarios and a phenomenon we term safety chain-of-thought CoT over-generalization, wherein safety-oriented reasoning patterns are excessively applied to benign queries, degrading helpfulness and user experience. To address these limitations, we propose Oyster-II, a reinforcement learning RL -based constructive safety alignment framework that adopts a Zero-RL paradigm combined with a multi-stage reinforcement learning strategy.Evaluated across extensive benchmarks, Oyster-II comprehensively surpasses both Qwen3-14B and its predecessor Oyster-I on safety dimensions, achieving cross-scale performance comparable to Qwen3-Max and Qwen3.5-397B.