PEA: Enhancing LLM Performance on Computational-Reasoning Tasks
This addresses the problem of enhancing LLM performance on computational-reasoning tasks for AI researchers, though it appears incremental as it builds on existing inference-time computation methods.
The study tackled the lack of a formal framework for characterizing reasoning tasks in LLMs by introducing the Predicate-Enumeration-Aggregation (PEA) framework, which decomposes computational reasoning problems and synthesizes programs, resulting in an average accuracy improvement of about 50% on benchmark tasks.
Large Language Models (LLMs) have exhibited remarkable capabilities across diverse domains, prompting investigations into their potential as generic reasoning engines. While recent studies have explored inference-time computation to enhance model performance on complex problems, current research lacks a formal framework to characterize the complexity of reasoning tasks. This study introduces the Predicate-Enumeration-Aggregation (PEA) framework, a formal approach to describe and solve a class of important reasoning tasks termed computational reasoning problems. The PEA framework decomposes these problems into predicate and enumeration components, using LLMs to synthesize programs based on specified predicates, enumeration, and aggregation rules. These synthesized programs are then executed to obtain solutions to the computational tasks. We demonstrate the framework's efficacy on benchmark tasks including Boolean satisfiability problems, game of $24$, and planning problems. Empirical evaluation reveals that PEA substantially enhances the performance of underlying models on benchmark computational problems, yielding an average accuracy improvement of approximately $50\%$, coupled with increased efficiency.