AEQ-Bench: Measuring Empathy of Omni-Modal Large Models
This addresses the problem of empathy assessment in AI for researchers and developers, but it is incremental as it builds on existing multimodal evaluation benchmarks.
The paper tackled the challenge of automatically evaluating empathy in omni-modal large models by introducing AEQ-Bench, a benchmark for assessing empathetic capabilities from audio and text inputs, finding that models with audio output outperformed text-only ones and aligned with human judgments for coarse-grained but not fine-grained aspects.
While the automatic evaluation of omni-modal large models (OLMs) is essential, assessing empathy remains a significant challenge due to its inherent affectivity. To investigate this challenge, we introduce AEQ-Bench (Audio Empathy Quotient Benchmark), a novel benchmark to systematically assess two core empathetic capabilities of OLMs: (i) generating empathetic responses by comprehending affective cues from multi-modal inputs (audio + text), and (ii) judging the empathy of audio responses without relying on text transcription. Compared to existing benchmarks, AEQ-Bench incorporates two novel settings that vary in context specificity and speech tone. Comprehensive assessment across linguistic and paralinguistic metrics reveals that (1) OLMs trained with audio output capabilities generally outperformed models with text-only outputs, and (2) while OLMs align with human judgments for coarse-grained quality assessment, they remain unreliable for evaluating fine-grained paralinguistic expressiveness.