In plain words: A set of puzzle-like planning tasks, drawn from classic action-and-change problems, tests whether language models can truly plan instead of just recalling answers. Even the strongest models fall well short on basic plan generation.
Abstract · PlanBench: An Extensible Benchmark for Evaluating Large Language Models on Planning and Reasoning about Change
Generating plans of action, and reasoning about change have long been considered a core competence of intelligent agents. It is thus no surprise that evaluating the planning and reasoning capabilities of large language models (LLMs) has become a hot topic of research. Most claims about LLM planning capabilities are however based on common sense tasks-where it becomes hard to tell whether LLMs are planning or merely retrieving from their vast world knowledge. There is a strong need for systematic and extensible planning benchmarks with sufficient diversity to evaluate whether LLMs have innate planning capabilities. Motivated by this, we propose PlanBench, an extensible benchmark suite based on the kinds of domains used in the automated planning community, especially in the International Planning Competition, to test the capabilities of LLMs in planning or reasoning about actions and change. PlanBench provides sufficient diversity in both the task domains and the specific planning capabilities. Our studies also show that on many critical capabilities-including plan generation-LLM performance falls quite short, even with the SOTA models. PlanBench can thus function as a useful marker of progress of LLMs in planning and reasoning.
Karthik Valmeekam, Matthew Marquez, Alberto Olmo, Sarath Sreedharan, Subbarao Kambhampati
arXiv:2206.10498 · cs.CL, cs.AI · submitted Jun 21, 2022 · updated Nov 26, 2023
abstract · pdf · html · NeurIPS 2023 Track on Datasets and Benchmarks
