用数学推理数据训练的检索模型，显著提升复杂任务表现。
RaDeR: Reasoning-aware Dense Retrieval Models
- 基于大模型推理轨迹生成多样化难样本，强化推理相关性训练
- 在Math和Coding任务上大幅超越基线，首次在思维链查询中胜过BM25
- 仅需2.5%数据量即达竞品性能，适合高效训练推理增强系统
我们提出RaDeR，一种基于数学问题求解过程中生成的数据训练的推理感知密集检索模型。该方法利用大语言模型的检索增强型推理路径与自反思相关性评估，生成多样且具有挑战性的推理相关性样本。训练后的RaDeR检索器在数学推理任务上表现出色，有效泛化至BRIGHT和RAR-b基准中的多种推理任务，整体性能持续优于强基线。特别地，在Math和Coding子集上表现显著更优。此外，RaDeR是首个在链式思维（Chain-of-Thought）查询下超越BM25的密集检索模型，凸显了推理驱动检索对增强语言模型的关键作用。同时，仅使用竞品REASONIR 2.5%的训练数据即达到相当或更优性能，证明了合成训练数据的质量优势。
原文摘要 · Abstract (English)
We propose RaDeR, a set of reasoning-based dense retrieval models trained with data derived from mathematical problem solving using large language models (LLMs). Our method leverages retrieval-augmented reasoning trajectories of an LLM and self-reflective relevance evaluation, enabling the creation of both diverse and hard-negative samples for reasoning-intensive relevance. RaDeR retrievers, trained for mathematical reasoning, effectively generalize to diverse reasoning tasks in the BRIGHT and RAR-b benchmarks, consistently outperforming strong baselines in overall performance. Notably, RaDeR achieves significantly higher performance than baselines on the Math and Coding splits. In addition, RaDeR presents the first dense retriever that outperforms BM25 when queries are Chain-of-Thought reasoning steps, underscoring the critical role of reasoning-based retrieval to augment reasoning language models. Furthermore, RaDeR achieves comparable or superior performance while using only 2.5% of the training data used by the concurrent work REASONIR, highlighting the quality of our synthesized training data.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。