November 11, 2025InformationOpen Access

Performance Evaluation Metrics for Empathetic LLMs

View Full Paper

Ask AI

Bookmark

Authors

Discussion

This framework measures emotional alignment in LLMs using sentiment and emotion metrics, suggesting improvements for AI interactions.

The new evaluation framework captures the intensity of empathy in LLM-generated responses, facilitating improved emotional interaction.
The proposed metric assesses emotional expression through specific components, such as Sentlink and NEmpathySort for sentiment.
This framework incorporates fine-grained emotional correspondence via Emosight, enhancing our understanding of LLM responses.
The semantic assessment based on RAGAS ensures contextual relevance and coherence, potentially improving conversational AI.

Hong et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark