@inproceedings{hu-etal-2025-context,
    title = "Context Reasoner: Incentivizing Reasoning Capability for Contextualized Privacy and Safety Compliance via Reinforcement Learning",
    author = "Hu, Wenbin  and
      Li, Haoran  and
      Jing, Huihao  and
      Hu, Qi  and
      Zeng, Ziqian  and
      Han, Sirui  and
      Heli, Xu  and
      Chu, Tianshu  and
      Hu, Peizhao  and
      Song, Yangqiu",
    editor = "Christodoulopoulos, Christos  and
      Chakraborty, Tanmoy  and
      Rose, Carolyn  and
      Peng, Violet",
    booktitle = "Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing",
    month = nov,
    year = "2025",
    address = "Suzhou, China",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-emnlp/2025.emnlp-main.44/",
    pages = "865--883",
    ISBN = "979-8-89176-332-6",
    abstract = "While Large Language Models (LLMs) exhibit remarkable capabilities, they also introduce significant safety and privacy risks. Current mitigation strategies often fail to preserve contextual reasoning capabilities in risky scenarios. Instead, they rely heavily on sensitive pattern matching to protect LLMs, which limits the scope. Furthermore, they overlook established safety and privacy standards, leading to systemic risks for legal compliance. To address these gaps, we formulate safety and privacy issues into contextualized compliance problems following the Contextual Integrity (CI) theory. Under the CI framework, we align our model with three critical regulatory standards: GDPR, EU AI Act, and HIPAA. Specifically, we employ reinforcement learning (RL) with a rule-based reward to incentivize contextual reasoning capabilities while enhancing compliance with safety and privacy norms. Through extensive experiments, we demonstrate that our method not only significantly enhances legal compliance (achieving a +8.58{\%} accuracy improvement in safety/privacy benchmarks) but also further improves general reasoning capability. For OpenThinker-7B, a strong reasoning model that significantly outperforms its base model Qwen2.5-7B-Instruct across diverse subjects, our method enhances its general reasoning capabilities, with +2.05{\%} and +8.98{\%} accuracy improvement on the MMLU and LegalBench benchmark, respectively."
}