LLMs have also been explored as zero-shot human models for boosting human-robot conversation. The research in [28] demonstrates that LLMs, educated on large textual content facts, can serve as helpful human models for specific HRI responsibilities, obtaining predictive performance comparabl