Embodied Agent Interface: Benchmarking LLMs for Embodied Decision Making

An LLM agent benchmark for embodied decision making: goal interpretation, subgoal decomposition, action sequencing, transition modeling. NeurIPS 2024 Oral.

关键词：Embodied Agent Interface, EAI, LLM agent benchmark, embodied decision making, embodied AI, LLM evaluation, goal interpretation, subgoal decomposition, action sequencing, transition modeling, VirtualH…