道德一致性流水线：面向大语言模型的持续道德评估
The Moral Consistency Pipeline: Continuous Ethical Evaluation for Large Language Models
AI总结：
MoCoP通过闭环框架持续评估大语言模型的道德一致性，揭示伦理与毒性间的强负相关，推动自主AI系统中计算道德研究。
摘要：
大语言模型（LLMs）的快速进步和适应性突显了道德一致性的重要性，即在不同情境中保持伦理一致的推理能力。现有的对齐框架，即旨在使模型行为与人类伦理和社会规范一致的结构化方法，通常依赖静态数据集和事后评估，难以深入理解伦理推理在不同情境或时间尺度下的演变。本研究提出了道德一致性流水线（MoCoP），一种无需数据集的闭环框架，用于持续评估和解释LLMs的道德稳定性。MoCoP结合了三个支持层：（i）词汇完整性分析，（ii）语义风险估计，以及（iii）基于推理的判断建模，这些层在自维持架构中协同工作，能够自主生成、评估和优化伦理情境，而无需外部监督。我们在GPT-4-Turbo和DeepSeek上的实验证实，MoCoP能够有效捕捉长期的伦理行为，揭示伦理维度与毒性维度之间存在显著的负相关关系（相关系数rET = -0.81，p值小于0.001），与响应延迟几乎无关联（相关系数rEL约为0）。这些发现表明，道德一致性和语言安全性更倾向于作为模型行为的稳定且可解释的特征，而非短期波动。此外，通过将伦理评估重新定义为一种动态且模型无关的道德内省形式，MoCoP提供了一个可重复的基础，用于可扩展的持续审计，并推动了自主AI系统中计算道德研究的发展。
英文摘要：
The rapid advancement and adaptability of Large Language Models (LLMs) highlight the need for moral consistency, the capacity to maintain ethically coherent reasoning across varied contexts. Existing alignment frameworks, structured approaches designed to align model behavior with human ethical and social norms, often rely on static datasets and post-hoc evaluations, offering limited insight into how ethical reasoning may evolve across different contexts or temporal scales. This study presents the Moral Consistency Pipeline (MoCoP), a dataset-free, closed-loop framework for continuously evaluating and interpreting the moral stability of LLMs. MoCoP combines three supporting layers: (i) lexical integrity analysis, (ii) semantic risk estimation, and (iii) reasoning-based judgment modeling within a self-sustaining architecture that autonomously generates, evaluates, and refines ethical scenarios without external supervision. Our empirical results on GPT-4-Turbo and DeepSeek suggest that MoCoP effectively captures longitudinal ethical behavior, revealing a strong inverse relationship between ethical and toxicity dimensions (correlation rET = -0.81, p value less than 0.001) and a near-zero association with response latency (correlation rEL approximately equal to 0). These findings demonstrate that moral coherence and linguistic safety tend to emerge as stable and interpretable characteristics of model behavior rather than short-term fluctuations. Furthermore, by reframing ethical evaluation as a dynamic, model-agnostic form of moral introspection, MoCoP offers a reproducible foundation for scalable, continuous auditing and advances the study of computational morality in autonomous AI systems.
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。
