pith. sign in

arxiv: 2401.00812 · v2 · pith:XZ6FRHMKnew · submitted 2024-01-01 · 💻 cs.CL

If LLM Is the Wizard, Then Code Is the Wand: A Survey on How Code Empowers Large Language Models to Serve as Intelligent Agents

classification 💻 cs.CL
keywords codellmslanguagemodelsabilityagentsexecutionfeedback
0
0 comments X
read the original abstract

The prominent large language models (LLMs) of today differ from past language models not only in size, but also in the fact that they are trained on a combination of natural language and formal language (code). As a medium between humans and computers, code translates high-level goals into executable steps, featuring standard syntax, logical consistency, abstraction, and modularity. In this survey, we present an overview of the various benefits of integrating code into LLMs' training data. Specifically, beyond enhancing LLMs in code generation, we observe that these unique properties of code help (i) unlock the reasoning ability of LLMs, enabling their applications to a range of more complex natural language tasks; (ii) steer LLMs to produce structured and precise intermediate steps, which can then be connected to external execution ends through function calls; and (iii) take advantage of code compilation and execution environment, which also provides diverse feedback for model improvement. In addition, we trace how these profound capabilities of LLMs, brought by code, have led to their emergence as intelligent agents (IAs) in situations where the ability to understand instructions, decompose goals, plan and execute actions, and refine from feedback are crucial to their success on downstream tasks. Finally, we present several key challenges and future directions of empowering LLMs with code.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. The Moltbook Files: A Harmless Slopocalypse or Humanity's Last Experiment

    cs.CL 2026-05 unverdicted novelty 7.0

    An AI-agent social platform generated mostly neutral content whose use in fine-tuning reduced model truthfulness comparably to human Reddit data, suggesting limited unique harm but flagging tail risks like secret leaks.

  2. Chat2Workflow: A Benchmark for Generating Executable Visual Workflows with Natural Language

    cs.CL 2026-04 unverdicted novelty 7.0

    Chat2Workflow benchmark shows that state-of-the-art LLMs often grasp high-level intent for visual workflow generation but fail to produce correct, stable, executable outputs, with an agentic framework delivering only ...

  3. NESA: Relational Neuro-Symbolic Static Program Analysis

    cs.PL 2024-12 conditional novelty 7.0

    NESA presents a neuro-symbolic framework that decomposes static analyses into policy-defined sub-problems solved by parsers and LLMs to enable compilation-free customizable analysis with reduced hallucinations.