EiCAP: Beyond Fluency, Probing and Improving Emotional Intelligence in LLMs via Psychologically Grounded Multi-Turn Dialogue
arXiv:2508.06196v2 Announce Type: replace Abstract: Large Language Models increasingly serve in emotionally sensitive roles, including mental health support, education, and crisis response, yet they lack a principled framework for assessing or improving Emotional Intelligence (EI). We introduce EiCAP, a unified, psychologically grounded six-layer EI taxonomy operationalized into two complementary resources. EiCAP-Bench is a multi-turn, one-vs-three forced-choice evaluation suite with 3,174 probes across 24 subcategories and cross-turn dependencies that reflect real conversational EI demands. E
As LLMs increasingly permeate sensitive applications, the immediate necessity to rigorously assess and improve their emotional intelligence (EI) becomes paramount for safe and effective deployment.
A principled framework for LLM emotional intelligence is critical for their responsible integration into human-centric roles, mitigating risks and unlocking new capabilities in mental health, education, and crisis support.
The introduction of EiCAP provides a standardized, psychologically grounded method to benchmark and enhance emotional intelligence in LLMs, shifting from ad-hoc assessments to structured improvement.
- · AI developers
- · Mental health support services
- · Educational technology providers
- · Crisis response organizations
- · Developers of emotionally naive LLMs
- · AI ethicists without specific EI assessment tools
The ability of LLMs to engage in more empathetic and contextually appropriate multi-turn dialogues will improve significantly.
Public trust and adoption of AI systems in sensitive domains will increase as their emotional intelligence becomes measurable and improvable.
The development of truly 'agentic' AI systems capable of complex social interactions and emotional reasoning could accelerate, profoundly impacting human-AI collaboration.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.CL