Benchmarks / AEQ
The Emotional Systems Quotient is a composite of the EQ-60 and SQ-R scores. Specifically AEQ = EQ-60 - SQ-R. Higher scores are better and if mapped to humans would statistically be typical of females.
The score is based on Empathizing-Systems Theory.
Most raw LLMs have an AEQ close to 0, i.e. their emotional capabilities are balanced out by their systemized thinking capabilities. In general, this is probably good in that LLM's need to be good at many things. In practice, this means that raw LLMs will need to be prompted or tuned with respect to identifying and manifesting emotions. Based on developing and testing generative AI for empathy, my experience is that typical prompting for emotional content results in shallow or repetitive responses that frequency drop out of empathetic mode or may feel empathetic in isolation but fail given a bigger context. In short, they will decay to this: It's Not About The Nail. This applies to all four CBT based therapy apps and 20 ChatGPT/CharacterAI bots I have tested.