Skip to content
OpenTrain AIFor AI Companies

HFEPX · Eval paper review

Multimodal Evaluator Preference Collapse: Cross-Modal Coupling in Self-Evolving Agents

Zewen Liu

Published

Jun 15, 2026

Citations

0

Trust level

Moderate

Usefulness score

40/100 (Low)

Extraction confidence

50% (Moderate)

Derived from extracted protocol signals and abstract evidence.

Rater population

Not reported

Signals refreshed

Jun 26, 2026

Open arXiv Find implementation Shortlist (0) RSS feed

Tags

Coding Pairwise Preference Tool Use

Should you rely on this paper?

This paper is adjacent to HFEPX scope and is best used for background context, not as a primary protocol reference.

Use this for comparison and orientation, not as your only source.

Best use

Background context only

Use if you need

A secondary eval reference to pair with stronger protocol papers.

What to verify

Read the full paper before copying any benchmark, metric, or protocol choices.

Main weakness

The abstract does not clearly name benchmarks or metrics.

Human feedback signal
Detected
From extracted signals
Evaluation signal
Detected
Eval setup described
Usefulness for eval research
40/100
Adjacent candidate

Treat as adjacent context, not a core eval-method reference.

If you are doing eval pipeline work, start here

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

Abstract

When AI agents use language models to evaluate their own outputs in a feedback loop, systematic biases emerge. We show that Evaluator Preference Collapse (EPC) is dramatically amplified in multimodal settings. Using GPT-4o to evaluate DeepSeek-chat across text and visual tasks, we find that a single strategy (step_by_step) absorbs 48.4% of all weight -- 3.2x the collapse observed in text-only self-evaluation -- while three visual-domain strategies receive only 9.1% combined weight. We then demonstrate a novel phenomenon we term cross-modal coupling: evaluator preferences acquired on one modality transfer to and corrupt strategy selection on another. Through a four-phase isolation training paradigm, we measure coupling coefficients and document strategy inversion -- the optimal strategy for a modality reverses after cross-modal exposure. A Phase 3 statistical validation across five evaluator configurations (N=80 total independent repetitions, ~35,000 API calls) with both text-proxy and real-image visual tasks finds: cross-model evaluation produces strong coupling (JSD~0.19-0.34), real-image inputs yield the most directionally consistent signal (mean gamma_{T->V}=1.145, gamma_{V->T}=0.937, 70% T->V, Cohen's d=0.56), and self-evaluation provides near-complete immunity -- 97% of runs (N=30) yield zero coupling (JSD=0.003, d=0.07). Three methodological ablations and multi-executor validation confirm the effect is not a structural artifact. We introduce the coupling matrix indexed by evaluator identity, release the MM-EPC framework, and identify cross-model evaluator architecture as the primary risk factor for preference drift. Code and data: https://github.com/aidless/mm-epc.

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.

"When AI agents use language models to evaluate their own outputs in a feedback loop, systematic biases emerge."

Evaluation Modes

missing

None explicit

Validate eval design from full paper text.

"When AI agents use language models to evaluate their own outputs in a feedback loop, systematic biases emerge."

Quality Controls

missing

Not reported

No explicit QC controls found.

"When AI agents use language models to evaluate their own outputs in a feedback loop, systematic biases emerge."

Benchmarks / Datasets

missing

Not extracted

No benchmark anchors detected.

"When AI agents use language models to evaluate their own outputs in a feedback loop, systematic biases emerge."

Reported Metrics

missing

Not extracted

No metric anchors detected.

"When AI agents use language models to evaluate their own outputs in a feedback loop, systematic biases emerge."

Benchmarks and datasets

No benchmark or dataset names were extracted from the available abstract.

Reported metrics

No metric terms were extracted from the available abstract.

Human feedback details
Uses human feedback
Yes
Feedback types
Pairwise Preference
Rater population
Not reported
Expertise required
Coding
Evaluation details
Evaluation modes
None
Agentic eval
Tool Use
Quality controls
Not reported
Evidence quality
Moderate
Use this page as
Background context only

Research brief

Metadata summary

When AI agents use language models to evaluate their own outputs in a feedback loop, systematic biases emerge.

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

Key takeaways

  • When AI agents use language models to evaluate their own outputs in a feedback loop, systematic biases emerge.
  • We show that Evaluator Preference Collapse (EPC) is dramatically amplified in multimodal settings.
  • Using GPT-4o to evaluate DeepSeek-chat across text and visual tasks, we find that a single strategy (step_by_step) absorbs 48.4% of all weight -- 3.2x the collapse observed in text-only self-evaluation -- while three visual-domain strategies receive only 9.1% combined weight.

Researcher actions

  • Compare this paper against nearby papers in the same arXiv category before using it for protocol decisions.
  • Validate inferred eval signals (Tool-use evaluation) 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

Contribution summary

  • We show that Evaluator Preference Collapse (EPC) is dramatically amplified in multimodal settings.
  • Using GPT-4o to evaluate DeepSeek-chat across text and visual tasks, we find that a single strategy (step_by_step) absorbs 48.4% of all weight -- 3.2x the collapse observed in text-only self-evaluation -- while three visual-domain…
  • We introduce the contagion matrix indexed by evaluator identity, release the MM-EPC framework, and identify cross-model evaluator architecture as the primary risk factor for preference drift.

Why it matters for eval

  • We show that Evaluator Preference Collapse (EPC) is dramatically amplified in multimodal settings.
  • We introduce the contagion matrix indexed by evaluator identity, release the MM-EPC framework, and identify cross-model evaluator architecture as the primary risk factor for preference drift.

Researcher checklist

  • Human feedback protocol is explicit

    Detected: Pairwise Preference

  • Evaluation mode is explicit

    No clear evaluation mode extracted.

  • Quality control reporting appears

    No calibration/adjudication/IAA control explicitly detected.

  • Benchmark or dataset anchors are present

    No benchmark/dataset anchor extracted from abstract.

  • Metric reporting is present

    No metric terms extracted.