Evaluating Position Bias in Large Language Model Recommendations

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Bito, Ethan, Ren, Yongli, He, Estrid
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916879126233088
author Bito, Ethan
Ren, Yongli
He, Estrid
author_facet Bito, Ethan
Ren, Yongli
He, Estrid
contents Large Language Models (LLMs) are being increasingly explored as general-purpose tools for recommendation tasks, enabling zero-shot and instruction-following capabilities without the need for task-specific training. While the research community is enthusiastically embracing LLMs, there are important caveats to directly adapting them for recommendation tasks. In this paper, we show that LLM-based recommendation models suffer from position bias, where the order of candidate items in a prompt can disproportionately influence the recommendations produced by LLMs. First, we analyse the position bias of LLM-based recommendations on real-world datasets, where results uncover systemic biases of LLMs with high sensitivity to input orders. Furthermore, we introduce a new prompting strategy to mitigate the position bias of LLM recommendation models called Ranking via Iterative SElection (RISE). We compare our proposed method against various baselines on key benchmark datasets. Experiment results show that our method reduces sensitivity to input ordering and improves stability without requiring model fine-tuning or post-processing.
format Preprint
id arxiv_https___arxiv_org_abs_2508_02020
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Evaluating Position Bias in Large Language Model Recommendations
Bito, Ethan
Ren, Yongli
He, Estrid
Information Retrieval
Large Language Models (LLMs) are being increasingly explored as general-purpose tools for recommendation tasks, enabling zero-shot and instruction-following capabilities without the need for task-specific training. While the research community is enthusiastically embracing LLMs, there are important caveats to directly adapting them for recommendation tasks. In this paper, we show that LLM-based recommendation models suffer from position bias, where the order of candidate items in a prompt can disproportionately influence the recommendations produced by LLMs. First, we analyse the position bias of LLM-based recommendations on real-world datasets, where results uncover systemic biases of LLMs with high sensitivity to input orders. Furthermore, we introduce a new prompting strategy to mitigate the position bias of LLM recommendation models called Ranking via Iterative SElection (RISE). We compare our proposed method against various baselines on key benchmark datasets. Experiment results show that our method reduces sensitivity to input ordering and improves stability without requiring model fine-tuning or post-processing.
title Evaluating Position Bias in Large Language Model Recommendations
topic Information Retrieval
url https://arxiv.org/abs/2508.02020