分阶段大模型流水线提升搜索相关性评估效率与精度
Multi-stage Large Language Model Pipelines Can Outperform GPT-4o in Relevance Assessment
- 将相关性评估拆解为多阶段模块化流程，用不同大小模型协同处理
- 在TREC-DL上比GPT-4o mini高出18.4%的Krippendorff's α，成本仅0.2美元/百万词元
- 适合需要低成本高可靠性的大规模搜索系统评测场景
搜索系统的有效性依赖于相关性标注，用于判断文档对特定查询和用户的有用程度。尽管来自真实用户的标注最理想，但规模化收集极为困难。因此常使用第三方标注员，但其标注一致性差，需投入大量成本进行审计、培训和监控。我们提出一种基于大语言模型的模块化分类流水线，将相关性评估任务分解为多个阶段，每个阶段采用不同的提示策略和不同规模、能力的模型。该方法应用于TREC Deep Learning（TREC-DL）数据集，其中一项方案在Krippendorff's α指标上较OpenAI的GPT-4o mini提升18.4%，同时保持约0.2美元/百万输入词元的成本，显著优于基准模型GPT-4o（5美元）。即使对GPT-4o旗舰版，该流水线也能将其α值提升9.7%。
原文摘要 · Abstract (English)
The effectiveness of search systems is evaluated using relevance labels that indicate the usefulness of documents for specific queries and users. While obtaining these relevance labels from real users is ideal, scaling such data collection is challenging. Consequently, third-party annotators are employed, but their inconsistent accuracy demands costly auditing, training, and monitoring. We propose an LLM-based modular classification pipeline that divides the relevance assessment task into multiple stages, each utilising different prompts and models of varying sizes and capabilities. Applied to TREC Deep Learning (TREC-DL), one of our approaches showed an 18.4% Krippendorff's $α$ accuracy increase over OpenAI's GPT-4o mini while maintaining a cost of about 0.2 USD per million input tokens, offering a more efficient and scalable solution for relevance assessment. This approach beats the baseline performance of GPT-4o (5 USD). With a pipeline approach, even the accuracy of the GPT-4o flagship model, measured in $α$, could be improved by 9.7%.
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。