面向安全图像生成的基于零样本智能体重写的价值对齐提示审核
Value-Aligned Prompt Moderation via Zero-Shot Agentic Rewriting for Safe Image Generation
AI总结：
针对文生图模型易生成不安全内容、现有防御难以兼顾价值对齐与生成质量的问题，提出零样本智能体框架VALOR，通过分层风险检测、提示重写与可选重生成，在保留创意的同时大幅降低不安全输出。
摘要：
Stable Diffusion等生成式视觉语言模型在创意媒体合成领域展现出卓越能力，但在遭遇对抗性提示时，也存在生成不安全、冒犯性或文化不当内容的重大风险。现有防御方案难以在不牺牲生成质量或不产生高昂成本的前提下，使输出与人类价值对齐。为应对这些挑战，我们提出了VALOR（价值对齐大语言模型监督重写器），这是一个模块化的零样本智能体框架，可实现更安全、更实用的文生图功能。VALOR将分层提示分析与人类对齐价值推理相结合：多级NSFW检测器过滤词汇与语义风险；文化价值对齐模块识别违反社会规范、法律规定与表征伦理的内容；意图消歧器检测隐蔽或间接的不安全暗示。当检测到不安全内容时，大语言模型会在动态、特定角色的指令下选择性重写提示，这些指令旨在保留用户意图的同时执行对齐约束。若生成的图像仍未通过安全检查，VALOR可选择性执行风格化重生成，在不改变核心语义的前提下将输出引导至更安全的视觉领域。在对抗性、模糊性和价值敏感提示上的实验表明，VALOR可将不安全输出最多降低100.00%，同时保留提示的实用性与创造性。这些结果表明VALOR是一种可扩展且有效的方案，可用于在开放世界场景中部署安全、对齐且实用的图像生成系统。
英文摘要：
Generative vision-language models like Stable Diffusion demonstrate remarkable capabilities in creative media synthesis, but they also pose substantial risks of producing unsafe, offensive, or culturally inappropriate content when prompted adversarially. Current defenses struggle to align outputs with human values without sacrificing generation quality or incurring high costs. To address these challenges, we introduce VALOR (Value-Aligned LLM-Overseen Rewriter), a modular, zero-shot agentic framework for safer and more helpful text-to-image generation. VALOR integrates layered prompt analysis with human-aligned value reasoning: a multi-level NSFW detector filters lexical and semantic risks; a cultural value alignment module identifies violations of social norms, legality, and representational ethics; and an intention disambiguator detects subtle or indirect unsafe implications. When unsafe content is detected, prompts are selectively rewritten by a large language model under dynamic, role-specific instructions designed to preserve user intent while enforcing alignment. If the generated image still fails a safety check, VALOR optionally performs a stylistic regeneration to steer the output toward a safer visual domain without altering core semantics. Experiments across adversarial, ambiguous, and value-sensitive prompts show that VALOR significantly reduces unsafe outputs by up to 100.00% while preserving prompt usefulness and creativity. These results highlight VALOR as a scalable and effective approach for deploying safe, aligned, and helpful image generation systems in open-world settings.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
