PoAct: Policy and Action Dual-Control Agent for Generalized Applications

TL;DR

PoAct is a novel dual-control agent framework that dynamically switches reasoning policies and modifies action spaces, significantly improving complex reasoning and code action quality in LLM-driven agents.

Contribution

This paper introduces PoAct, a dual-control agent that enhances reasoning accuracy and code actions by managing complex action spaces and policy switching, addressing limitations of existing frameworks.

Findings

20% improvement on LegalAgentBench

Reduces token consumption in complex tasks

Demonstrates scalability with GPT-4o and GLM-4 models

Abstract

Based on their superior comprehension and reasoning capabilities, Large Language Model (LLM) driven agent frameworks have achieved significant success in numerous complex reasoning tasks. ReAct-like agents can solve various intricate problems step-by-step through progressive planning and tool calls, iteratively optimizing new steps based on environmental feedback. However, as the planning capabilities of LLMs improve, the actions invoked by tool calls in ReAct-like frameworks often misalign with complex planning and challenging data organization. Code Action addresses these issues while also introducing the challenges of a more complex action space and more difficult action organization. To leverage Code Action and tackle the challenges of its complexity, this paper proposes Policy and Action Dual-Control Agent (PoAct) for generalized applications. The aim is to achieve higher-quality code actions and more accurate reasoning paths by dynamically switching reasoning policies and modifying the action space. Experimental results on the Agent Benchmark for both legal and generic scenarios demonstrate the superior reasoning capabilities and reduced token consumption of our approach in complex tasks. On the LegalAgentBench, our method shows a 20 percent improvement over the baseline while requiring fewer tokens. We conducted experiments and analyses on the GPT-4o and GLM-4 series models, demonstrating the significant potential and scalability of our approach to solve complex problems.