Reproduce “To Embody or Not: The Effect Of Embodiment On User Perception Of LLM-based Conversational Agents” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Kyra Wang, Abhinaba Roy, Boon-Kiat Quek, Jessica Goh, Dorien Herremans
The authors compared a human-like embodied agent with a text-only agent in a 20-person survival-planning pilot using the same LLaMA model. Mean competence ratings were 3.54 for the embodied agent and 4.26 for the text-only agent. Participants also reported that the embodied agent agreed too readily, but scenario differences confounded embodiment, and sycophancy was not independently measured.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.