Listwise Cross-Encoder Fine-Tuning vs. Agentic Instruction Tuning for LLM Rerankers: A Systematic Study in Medical Procedure Reranking
TL;DR
A systematic comparison of two reranking paradigms for this production task finds that a 109M-parameter cross-encoder fine-tuned with ListNet outperforms the 4B-parameter model by 2.6 percentage points on NDCG@3 and 13.3 points on Spearman correlation - at 37x fewer parameters.
Abstract
Reranking medical procedures against patient queries is a critical component of health insurance information retrieval, complicated by a substantial lexical gap between patient language and clinical nomenclature. We present a systematic comparison of two reranking paradigms for this production task: (1) small cross-encoders (MedCPT, MiniLM-L12) fine-tuned with listwise learning-to-rank objectives across layer freezing configurations, and (2) Qwen3-Reranker-4B, a 4B-parameter instruction reranker whose prompt is iteratively refined via an agentic optimization loop driven by GPT-4.1. On a purpose-built dataset of 2,647 queries across 708 insurance services, we find that a 109M-parameter cross-encoder fine-tuned with ListNet outperforms the 4B-parameter model by 2.6 percentage points on NDCG@3 and 13.3 points on Spearman correlation - at 37x fewer parameters. We report practical findings, a scalable LLM based dataset construction pipeline, and deployment trade-offs relevant to production reranking systems. We release our code and a sample dataset to support reproducibility and adaptation to other domains.