Skip to content
OpenTrain AIFor AI Companies

HFEPX · Eval paper review

Nine Judges, Two Effective Votes: Correlated Errors Undermine LLM Evaluation Panels

Guneet Kohli

Published

May 28, 2026

Citations

0

Trust level

High

Usefulness score

77/100 (High)

Extraction confidence

80% (High)

Derived from extracted protocol signals and abstract evidence.

Rater population

Not reported

Signals refreshed

May 28, 2026

Open arXiv Find implementation Shortlist (0) RSS feed

Tags

Automatic Metrics General Llm As Judge Pairwise Preference

Should you rely on this paper?

This paper has strong direct human-feedback and evaluation protocol signal and is suitable as a primary eval pipeline reference.

Use this as a practical starting point for protocol research, then validate against the original paper.

Best use

Primary benchmark and eval reference

Use if you need

A benchmark-and-metrics comparison anchor.

What to verify

Validate the evaluation procedure and quality controls in the full paper before operational use.

Main weakness

No major weakness surfaced.

Human feedback signal
Detected
From extracted signals
Evaluation signal
Detected
Eval setup described
Usefulness for eval research
77/100
High-confidence candidate

Use this as a primary source when designing or comparing eval protocols.

If you are doing eval pipeline work, start here

Human Eval Papers LLM-as-Judge Papers Pairwise Preference Papers Tool-Use Evaluation Search

Abstract

LLM-as-a-judge panels aggregate votes from multiple models, with the expectation that diverse models yield more reliable evaluations. We develop a framework to measure the true informational value of such panels and quantify how far their reliability falls short of the independent-voting ideal. Testing a panel of 9 frontier LLMs from 7 model families on three natural language inference datasets (each with 100 human annotations per item), we find that the 9 judges effectively provide only about 2 independent votes' worth of information. Roughly three-quarters of the panel's nominal independence is lost because the models make the same mistakes on the same items. The consequences are stark: the panel's actual accuracy falls 8-22 percentage points short of what independent voting would achieve, and the best single judge matches or outperforms the full panel across all conditions. Neither adding more judges nor using smarter aggregation algorithms helps -- established methods close at most 11% of this gap, even with access to the correct answers. We quantify these findings using the Kish effective sample size (n_eff) and a Condorcet null model, and show the deficit is robust across prompt variants, temperatures, chain-of-thought reasoning, and a pairwise preference task (RewardBench). The bottleneck is correlated judges, not the aggregation algorithm, implying that scaling up panels cannot substitute for genuinely independent evaluation.

What we could verify

These are the protocol signals we could actually recover from the available paper metadata. Use them to decide whether this paper is worth deeper reading.

Human Feedback Types

strong

Pairwise Preference

Directly usable for protocol triage.

"LLM-as-a-judge panels aggregate votes from multiple models, with the expectation that diverse models yield more reliable evaluations."

Evaluation Modes

strong

Llm As Judge, Automatic Metrics

Includes extracted eval setup.

"LLM-as-a-judge panels aggregate votes from multiple models, with the expectation that diverse models yield more reliable evaluations."

Quality Controls

missing

Not reported

No explicit QC controls found.

"LLM-as-a-judge panels aggregate votes from multiple models, with the expectation that diverse models yield more reliable evaluations."

Benchmarks / Datasets

strong

Rewardbench

Useful for quick benchmark comparison.

"We quantify these findings using the Kish effective sample size (n_eff) and a Condorcet null model, and show the deficit is robust across prompt variants, temperatures, chain-of-thought reasoning, and a pairwise preference task (RewardBench)."

Reported Metrics

strong

Accuracy

Useful for evaluation criteria comparison.

"The consequences are stark: the panel's actual accuracy falls 8-22 percentage points short of what independent voting would achieve, and the best single judge matches or outperforms the full panel across all conditions."

Benchmarks and datasets

Rewardbench

Reported metrics

accuracy
Human feedback details
Uses human feedback
Yes
Feedback types
Pairwise Preference
Rater population
Not reported
Unit of annotation
Pairwise
Expertise required
General
Evaluation details
Evaluation modes
Llm As Judge, Automatic Metrics
Agentic eval
None
Quality controls
Not reported
Evidence quality
High
Use this page as
Primary benchmark and eval reference

Research brief

Metadata summary

LLM-as-a-judge panels aggregate votes from multiple models, with the expectation that diverse models yield more reliable evaluations.

Based on abstract + metadata only. Check the source paper before making high-confidence protocol decisions.

Key takeaways

  • LLM-as-a-judge panels aggregate votes from multiple models, with the expectation that diverse models yield more reliable evaluations.
  • We develop a framework to measure the true informational value of such panels and quantify how far their reliability falls short of the independent-voting ideal.
  • Testing a panel of 9 frontier LLMs from 7 model families on three natural language inference datasets (each with 100 human annotations per item), we find that the 9 judges effectively provide only about 2 independent votes' worth of information.

Researcher actions

  • Compare this paper against nearby papers in the same arXiv category before using it for protocol decisions.
  • Validate inferred eval signals (Automatic metrics) against the full paper.
  • Use related-paper links to find stronger protocol-specific references.

Caveats

  • Generated from abstract + metadata only; no PDF parsing.
  • Signals below are heuristic and may miss details reported outside the abstract.

Recommended queries

Pairwise preference evaluation Accuracy

Contribution summary

  • We develop a framework to measure the true informational value of such panels and quantify how far their reliability falls short of the independent-voting ideal.
  • The consequences are stark: the panel's actual accuracy falls 8-22 percentage points short of what independent voting would achieve, and the best single judge matches or outperforms the full panel across all conditions.
  • Neither adding more judges nor using smarter aggregation algorithms helps -- established methods close at most 11% of this gap, even with access to the correct answers.

Why it matters for eval

  • The consequences are stark: the panel's actual accuracy falls 8-22 percentage points short of what independent voting would achieve, and the best single judge matches or outperforms the full panel across all conditions.
  • Neither adding more judges nor using smarter aggregation algorithms helps -- established methods close at most 11% of this gap, even with access to the correct answers.

Researcher checklist

  • Human feedback protocol is explicit

    Detected: Pairwise Preference

  • Evaluation mode is explicit

    Detected: Llm As Judge, Automatic Metrics

  • Quality control reporting appears

    No calibration/adjudication/IAA control explicitly detected.

  • Benchmark or dataset anchors are present

    Detected: Rewardbench

  • Metric reporting is present

    Detected: accuracy