Papers
arxiv:2512.09066

ORCA: Open-ended Response Correctness Assessment for Audio Question Answering

Published on Nov 28, 2025
Authors:
,
,
,
,
,
,
,
,
,
,
,

Abstract

A framework for evaluating open-ended responses from large audio language models that models human judgment variability using Beta distributions to predict correctness and uncertainty, achieving high correlation with human judgments while reducing computational costs.

Evaluating open-ended responses from large audio language models (LALMs) is challenging because human annotators often genuinely disagree on answer correctness due to multiple valid interpretations, partial correctness, and subjective judgment. Traditional metrics reporting only mean scores fail to capture this uncertainty. We present ORCA (Open-ended Response Correctness Assessment), a framework that models the variability in human judgments using Beta distributions to predict both expected correctness and uncertainty. Our three-stage annotation framework combines human judgment with structured feedback and iterative refinement to simultaneously curate training data and improve benchmark quality. We collected 11,721 annotations across 3,580 question-answer pairs from 15 LALMs on two audio QA benchmarks, achieving inter-annotator agreement of 0.82 (Krippendorff's alpha). ORCA achieves 0.91 Spearman correlation with mean human judgments, matching or outperforming LLM-judge baselines while providing uncertainty estimates and requiring significantly less compute. We release our models, code, and curated dataset.

Community

Sign up or log in to comment

Get this paper in your agent:

hf papers read 2512.09066
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2512.09066 in a model README.md to link it from this page.

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2512.09066 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2512.09066 in a Space README.md to link it from this page.

Collections including this paper 1