DiPT:通过多样化视角增强 LLM 推理
DiPT: Enhancing LLM reasoning through diversified perspective-taking
AI总结:
本文提出 DiPT 方法,通过在推理阶段融入多样化视角来补充单一推理路径,提升 LLM 的推理性能、稳定性及安全性,并提供以数据为中心的微调增强方案。
摘要:
现有关于改进语言模型推理的工作通常探索单一解决路径,这容易产生错误。受社会科学中视角采择的启发,本文引入了 DiPT,这是一种通过明确融入多样化视角来补充当前推理方法的新颖方法。这种方法允许模型在推理阶段更深入地理解问题背景,并识别出最有效的解决路径。此外,它提供了一种通用的以数据为中心的 AI 方案,用于增强现有数据以提高其用于微调的质量。我们的实证结果表明,DiPT 可以灵活地集成到专注于单一推理方法的现有方法中,在面对复述问题时增强其推理性能和稳定性。此外,我们通过在面对旨在绕过部署模型内置安全防护的“越狱”提示时维持模型的安全输出,阐明了其上下文理解能力的提升。最后,我们表明,与仅使用原始数据进行微调相比,使用富含多样化视角的数据进行微调可以提升模型的推理能力。
英文摘要:
Existing work on improving language model reasoning typically explores a single solution path, which can be prone to errors. Inspired by perspective-taking in social studies, this paper introduces DiPT, a novel approach that complements current reasoning methods by explicitly incorporating diversified viewpoints. This approach allows the model to gain a deeper understanding of the problem's context and identify the most effective solution path during the inference stage. Additionally, it provides a general data-centric AI recipe for augmenting existing data to improve their quality for fine-tuning. Our empirical results demonstrate that DiPT can be flexibly integrated into existing methods that focus on a single reasoning approach, enhancing their reasoning performance and stability when presented with paraphrased problems. Furthermore, we illustrate improved context understanding by maintaining the model's safe outputs against "jailbreaking" prompts intentionally designed to bypass safeguards built into deployed models. Lastly, we show that fine-tuning with data enriched with diverse perspectives can boost the reasoning capabilities of the model compared to fine-tuning with raw data alone.
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。