TAPE: Tool-Guided Adaptive Planning and Constrained Execution in Language Model Agents

TL;DR

TAPE introduces a tool-guided adaptive planning and constrained execution framework that significantly improves the success rates of language model agents in complex, strict environments by combining multi-plan graph aggregation, external solving, and adaptive re-planning.

Contribution

The paper presents TAPE, a novel framework that enhances language model agent robustness through graph-based planning, constrained decoding, and adaptive re-planning, addressing limitations of existing methods.

Findings

TAPE outperforms existing frameworks across multiple tasks.

Success rates increase by 21 percentage points on hard settings.

Significant improvements for weaker base models.

Abstract

Language Model (LM) agents have demonstrated remarkable capabilities in solving tasks that require multiple interactions with the environment. However, they remain vulnerable in environments where a single error often leads to irrecoverable failure, particularly under strict feasibility constraints. We systematically analyze existing agent frameworks, identifying imperfect planning and stochastic execution as the primary causes. To address these challenges, we propose Tool-guided Adaptive Planning with constrained Execution (TAPE). TAPE enhances planning capability by aggregating multiple plans into a graph and employing an external solver to identify a feasible path. During execution, TAPE employs constrained decoding to reduce sampling noise, while adaptively re-planning whenever environmental feedback deviates from the intended state. Experiments across Sokoban, ALFWorld, MuSiQue, and GSM8K-Hard demonstrate that TAPE consistently outperforms existing frameworks, with particularly large gains on hard settings, improving success rates by 21.0 percentage points on hard settings on average, and by 20.0 percentage points for weaker base models on average. Code and data available at here.