Saved in:
Bibliographic Details
Main Authors: Mansour, Watheq, Culpepper, J. Shane, Mackenzie, Joel, Yates, Andrew
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.05603
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911374407368704
author Mansour, Watheq
Culpepper, J. Shane
Mackenzie, Joel
Yates, Andrew
author_facet Mansour, Watheq
Culpepper, J. Shane
Mackenzie, Joel
Yates, Andrew
contents Using large language models (LLMs) to annotate relevance is an increasingly important technique in the information retrieval community. While some studies demonstrate that LLMs can achieve high user agreement with ground truth (human) judgments, other studies have argued for the opposite conclusion. To the best of our knowledge, these studies have primarily focused on classic ad-hoc text search scenarios. In this paper, we conduct an analysis on user agreement between LLM and human experts, and explore the impact disagreement has on system rankings. In contrast to prior studies, we focus on a collection composed of audio files that are transcribed into two-minute segments -- the TREC 2020 and 2021 podcast track. We employ five different LLM models to re-assess all of the query-segment pairs, which were originally annotated by TREC assessors. Furthermore, we re-assess a small subset of pairs where LLM and TREC assessors have the highest disagreement, and found that the human experts tend to agree with LLMs more than with the TREC assessors. Our results reinforce the previous insights of Sormunen in 2002 -- that relying on a single assessor leads to lower user agreement.
format Preprint
id arxiv_https___arxiv_org_abs_2601_05603
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Revisiting Human-vs-LLM judgments using the TREC Podcast Track
Mansour, Watheq
Culpepper, J. Shane
Mackenzie, Joel
Yates, Andrew
Information Retrieval
Using large language models (LLMs) to annotate relevance is an increasingly important technique in the information retrieval community. While some studies demonstrate that LLMs can achieve high user agreement with ground truth (human) judgments, other studies have argued for the opposite conclusion. To the best of our knowledge, these studies have primarily focused on classic ad-hoc text search scenarios. In this paper, we conduct an analysis on user agreement between LLM and human experts, and explore the impact disagreement has on system rankings. In contrast to prior studies, we focus on a collection composed of audio files that are transcribed into two-minute segments -- the TREC 2020 and 2021 podcast track. We employ five different LLM models to re-assess all of the query-segment pairs, which were originally annotated by TREC assessors. Furthermore, we re-assess a small subset of pairs where LLM and TREC assessors have the highest disagreement, and found that the human experts tend to agree with LLMs more than with the TREC assessors. Our results reinforce the previous insights of Sormunen in 2002 -- that relying on a single assessor leads to lower user agreement.
title Revisiting Human-vs-LLM judgments using the TREC Podcast Track
topic Information Retrieval
url https://arxiv.org/abs/2601.05603