# Multi-stage Large Language Model Pipelines Can Outperform GPT-4o in Relevance Assessment

arXiv: https://arxiv.org/abs/2501.14296 · HTML: https://mlsanity.com/paper/2501.14296/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 24 Jan 2025.

## Built on by (most cited first, 7 of 7)

- [Exploring Autonomous Agents: A Closer Look at Why They Fail When Completing Tasks](https://mlsanity.com/paper/2508.13143/) (2025)
- [Synthetic Data Generation for Phrase Break Prediction with Large Language Model](https://mlsanity.com/paper/2507.18044/) (2025)
- [Bridging Technology and Humanities: Evaluating the Impact of Large Language Models on Social Sciences Research with DeepSeek-R1](https://mlsanity.com/paper/2503.16304/) (2025)
- [Large Language Models in the Task of Automatic Validation of Text Classifier Predictions](https://mlsanity.com/paper/2505.18688/) (2025)
- [Re-Rankers as Relevance Judges](https://mlsanity.com/paper/2601.04455/) (2026)
- [HIERA: Hierarchical Multi-Agent Relevance Assessment for Content Discovery Systems](https://mlsanity.com/paper/2608.00785/) (2026)
- [When Do LLMs Actually Help? Evaluating LLMs as Data Quality Annotators](https://mlsanity.com/paper/2608.18158/) (2026)

## Abstract (arXiv)

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%.