Large Language Models as General Pattern Machines

Suvir Mirchandani, Fei Xia, Pete Florence, brian ichter, Danny Driess, Montserrat Gonzalez Arenas, Kanishka Rao, Dorsa Sadigh, Andy Zeng

Keywords: large language models, in-context learning, language for robotics

TL;DR: We explore the utility of large language models to function as "general pattern machines" for robotics and sequential decision making, enabled by their capability to perform in-context learning over abstract sequences.

Abstract: We observe that pre-trained large language models (LLMs) are capable of autoregressively completing complex token sequences—from arbitrary ones procedurally generated by probabilistic context-free grammars (PCFG), to more rich spatial patterns found in the Abstraction and Reasoning Corpus (ARC), a general AI benchmark, prompted in the style of ASCII art. Surprisingly, pattern completion proficiency can be partially retained even when the sequences are expressed using tokens randomly sampled from the vocabulary. These results suggest that without any additional training, LLMs can serve as general sequence modelers, driven by in-context learning. In this work, we investigate how these zero-shot capabilities may be applied to problems in robotics—from extrapolating sequences of numbers that represent states over time to complete simple motions, to least-to-most prompting of reward-conditioned trajectories that can discover and represent closed-loop policies (e.g., a stabilizing controller for CartPole). While difficult to deploy today for real systems due to latency, context size limitations, and compute costs, the approach of using LLMs to drive low-level control may provide an exciting glimpse into how the patterns among words could be transferred to actions.

Student First Author: yes

Instructions: I have read the instructions for authors (https://corl2023.org/instructions-for-authors/)

Website: https://general-pattern-machines.github.io