StateFlow: Enhancing LLM Task-Solving through State-Driven Workflows

TL;DR

StateFlow introduces a state machine-based paradigm for LLM task-solving, improving control, interpretability, and efficiency in complex, multi-step tasks involving external tools and dynamic interactions.

Contribution

It proposes a novel state-driven framework for LLMs that separates process grounding from sub-task solving, enhancing adaptability and performance.

Findings

Achieves 13% higher success rate on InterCode SQL benchmark

Achieves 28% higher success rate on ALFWorld benchmark

Reduces task-solving cost by up to 5x

Abstract

It is a notable trend to use Large Language Models (LLMs) to tackle complex tasks, e.g., tasks that require a sequence of actions and dynamic interaction with tools and external environments. In this paper, we propose StateFlow, a novel LLM-based task-solving paradigm that conceptualizes complex task-solving processes as state machines. In StateFlow, we distinguish between "process grounding" (via state and state transitions) and "sub-task solving" (through actions within a state), enhancing control and interpretability of the task-solving procedure. A state represents the status of a running process. The transitions between states are controlled by heuristic rules or decisions made by the LLM, allowing for a dynamic and adaptive progression. Upon entering a state, a series of actions is executed, involving not only calling LLMs guided by different prompts, but also the utilization of external tools as needed. Our results show that StateFlow significantly enhances LLMs' efficiency. For instance, StateFlow achieves 13% and 28% higher success rates compared to ReAct in InterCode SQL and ALFWorld benchmark, with 5x and 3x less cost respectively. We also show that StateFlow can be combined with iterative refining methods like Reflexion to further improve performance.