Preemptive Detection and Correction of Misaligned Actions in LLM Agents

TL;DR

This paper presents InferAct, a novel LLM-based approach that preemptively detects and corrects misaligned actions in agents, significantly reducing undesirable outcomes and improving reliability in real-world applications.

Contribution

InferAct introduces a belief reasoning method grounded in Theory-of-Mind for preemptive detection and correction of misaligned actions in LLM agents, a relatively unexplored area.

Findings

Achieves up to 20% improvement in misaligned action detection.

Effectively alerts users for timely correction.

Enhances overall agent alignment and reliability.

Abstract

Deploying LLM-based agents in real-life applications often faces a critical challenge: the misalignment between agents' behavior and user intent. Such misalignment may lead agents to unintentionally execute critical actions that carry negative outcomes (e.g., accidentally triggering a "buy-now" in web shopping), resulting in undesirable or even irreversible consequences. Although addressing these issues is crucial, the preemptive detection and correction of misaligned actions remains relatively underexplored. To fill this gap, we introduce InferAct, a novel approach that leverages the belief reasoning ability of LLMs, grounded in Theory-of-Mind, to detect misaligned actions before execution. Once the misalignment is detected, InferAct alerts users for timely correction, preventing adverse outcomes and enhancing the reliability of LLM agents' decision-making processes. Experiments on three widely used tasks demonstrate that InferAct achieves up to 20% improvements on Marco-F1 against baselines in misaligned action detection. An in-depth evaluation of misalignment correction further highlights InferAct's effectiveness in improving agent alignment.