View in EDS

Probing subjective judgment variance in LLM evaluators: A framework for robust IR evaluation.

Saved in:
Bibliographic Details
Title: Probing subjective judgment variance in LLM evaluators: A framework for robust IR evaluation.
Authors: Kumar, Ritesh1 (AUTHOR) riteshkumar@iiitsurat.ac.in
Source: Applied Intelligence. Apr2026, Vol. 56 Issue 6, p1-22. 22p.
Abstract: Large Language Models (LLMs) are increasingly used as evaluators for subjective Information Retrieval (IR) tasks, yet their reliability under ambiguity remains poorly understood. We systematically probe LLM-as-a-judge systems by deliberately maximizing inter-model disagreement, which we formalize as horizontal variance. Across five widely used LLM evaluators, we observe that ambiguity-based prompts induce up to a 63% increase in horizontal variance, while intra-model (vertical) variance remains stable. This indicates systematic, not random, evaluator inconsistency. Our results show that models may appear stable in isolation yet diverge substantially when evaluated collectively, exposing hidden bias and instability. These findings highlight the risks of naive deployment of LLM evaluators in fairness-sensitive IR pipelines and motivate multi-model, variance-aware evaluation strategies. [ABSTRACT FROM AUTHOR]
Copyright of Applied Intelligence is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Engineering Source
FullText Text:
  Availability: 0
Header DbId: egs
DbLabel: Engineering Source
An: 193385537
AccessLevel: 6
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Probing subjective judgment variance in LLM evaluators: A framework for robust IR evaluation.
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Kumar%2C+Ritesh%22">Kumar, Ritesh</searchLink><relatesTo>1</relatesTo> (AUTHOR)<i> riteshkumar@iiitsurat.ac.in</i>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="JN" term="%22Applied+Intelligence%22">Applied Intelligence</searchLink>. Apr2026, Vol. 56 Issue 6, p1-22. 22p.
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Large Language Models (LLMs) are increasingly used as evaluators for subjective Information Retrieval (IR) tasks, yet their reliability under ambiguity remains poorly understood. We systematically probe LLM-as-a-judge systems by deliberately maximizing inter-model disagreement, which we formalize as horizontal variance. Across five widely used LLM evaluators, we observe that ambiguity-based prompts induce up to a 63% increase in horizontal variance, while intra-model (vertical) variance remains stable. This indicates systematic, not random, evaluator inconsistency. Our results show that models may appear stable in isolation yet diverge substantially when evaluated collectively, exposing hidden bias and instability. These findings highlight the risks of naive deployment of LLM evaluators in fairness-sensitive IR pipelines and motivate multi-model, variance-aware evaluation strategies. [ABSTRACT FROM AUTHOR]
– Name: AbstractSuppliedCopyright
  Label:
  Group: Ab
  Data: <i>Copyright of Applied Intelligence is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract.</i> (Copyright applies to all Abstracts.)
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=egs&AN=193385537
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1007/s10489-026-07244-8
    Languages:
      – Code: eng
        Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 22
        StartPage: 1
    Titles:
      – TitleFull: Probing subjective judgment variance in LLM evaluators: A framework for robust IR evaluation.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Kumar, Ritesh
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 15
              M: 04
              Text: Apr2026
              Type: published
              Y: 2026
          Identifiers:
            – Type: issn-print
              Value: 0924669X
          Numbering:
            – Type: volume
              Value: 56
            – Type: issue
              Value: 6
          Titles:
            – TitleFull: Applied Intelligence
              Type: main
ResultId 1