微调后的安全性漂移：来自高风险领域的证据
Safety Drift After Fine-Tuning: Evidence from High-Stakes Domains
AI总结：
研究发现微调导致的安全性变化显著且不一致，挑战了基于基础模型的安全性假设，需重新评估微调模型以管理下游风险。
摘要：
基础模型通常在特定领域进行微调，但安全性评估通常仅针对基础模型，隐含假设安全性在下游适应中保持不变。我们通过分析100个模型的安全行为，包括医疗和法律领域广泛应用的微调模型及开放基础模型的受控适应，发现良性微调导致测量安全性出现大规模、异质且常矛盾的变化：模型在某些指标上改进而在其他指标上退化，评估结果存在显著分歧。这些结果表明，普通下游适应下安全性不稳定，引发关于基于基础模型评估的治理和部署实践的关键问题。在部署相关情境中缺乏对微调模型的显式重新评估，此类方法无法充分管理下游风险，忽视了实际危害来源——在高风险领域这些失败尤其具有重大影响，挑战了当前的问责范式。
英文摘要：
Foundation models are routinely fine-tuned for use in particular domains, yet safety assessments are typically conducted only on base models, implicitly assuming that safety properties persist through downstream adaptation. We test this assumption by analyzing the safety behavior of 100 models, including widely deployed fine-tunes in the medical and legal domains as well as controlled adaptations of open foundation models alongside their bases. Across general-purpose and domain-specific safety benchmarks, we find that benign fine-tuning induces large, heterogeneous, and often contradictory changes in measured safety: models frequently improve on some instruments while degrading on others, with substantial disagreement across evaluations. These results show that safety behavior is not stable under ordinary downstream adaptation, raising critical questions about governance and deployment practices centered on base-model evaluations. Without explicit re-evaluation of fine-tuned models in deployment-relevant contexts, such approaches fall short of adequately managing downstream risk, overlooking practical sources of harm -- failures that are especially consequential in high-stakes settings and challenge current accountability paradigms.
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。
