A Fast, Reliable, and Secure Programming Language for LLM Agents with Code Actions

TL;DR

The paper introduces Quasar, a new programming language designed for LLM agents to perform code actions more securely, reliably, and efficiently, with automatic parallelization and uncertainty quantification.

Contribution

It presents Quasar, a novel language for code actions that enhances performance, security, and reliability of LLM agents compared to using Python.

Findings

Reduced execution time by 42%

Decreased user approval interactions by 52%

Achieved targeted reliability with conformal prediction

Abstract

Modern large language models (LLMs) are often deployed as agents, calling external tools adaptively to solve tasks. Rather than directly calling tools, it can be more effective for LLMs to write code to perform the tool calls, enabling them to automatically generate complex control flow such as conditionals and loops. Such code actions are typically provided as Python code, since LLMs are quite proficient at it; however, Python may not be the ideal language due to limited built-in support for performance, security, and reliability. We propose a novel programming language for code actions, called Quasar, which has several benefits: (1) automated parallelization to improve performance, (2) uncertainty quantification to improve reliability and mitigate hallucinations, and (3) security features enabling the user to validate actions. LLMs can write code in a subset of Python, which is automatically transpiled to Quasar. We evaluate our approach on the ViperGPT visual question answering agent, applied to the GQA dataset, demonstrating that LLMs with Quasar actions instead of Python actions retain strong performance, while reducing execution time when possible by 42%, improving security by reducing user approval interactions when possible by 52%, and improving reliability by applying conformal prediction to achieve a desired target coverage level.