Guiding LLM Temporal Logic Generation with Explicit Separation of Data and Control

TL;DR

This paper investigates how explicitly separating control and data guidance improves Large Language Models' ability to generate temporal logic specifications for reactive system synthesis, providing a benchmark for future evaluation.

Contribution

It introduces a guidance method that separates control and data in LLM prompts, enhancing temporal logic specification generation and providing a benchmark for future research.

Findings

Separation of control and data improves specification quality

Benchmark set enables standardized evaluation of LLMs

Guidance method enhances accessibility of temporal logic writing

Abstract

Temporal logics are powerful tools that are widely used for the synthesis and verification of reactive systems. The recent progress on Large Language Models (LLMs) has the potential to make the process of writing such specifications more accessible. However, writing specifications in temporal logics remains challenging for all but the most expert users. A key question in using LLMs for temporal logic specification engineering is to understand what kind of guidance is most helpful to the LLM and the users to easily produce specifications. Looking specifically at the problem of reactive program synthesis, we explore the impact of providing an LLM with guidance on the separation of control and data--making explicit for the LLM what functionality is relevant for the specification, and treating the remaining functionality as an implementation detail for a series of pre-defined functions and predicates. We present a benchmark set and find that this separation of concerns improves specification generation. Our benchmark provides a test set against which to verify future work in LLM generation of temporal logic specifications.