# NL in the Middle: Code Translation with LLMs and Intermediate Representations

arXiv: https://arxiv.org/abs/2507.08627 · HTML: https://mlsanity.com/paper/2507.08627/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 11 Jul 2025.

## Built on by (most cited first, 3 of 3)

- [Specification-Based Code-Text-Code Reengineering for LLM-Mediated Software Evolution](https://mlsanity.com/paper/2605.25232/) (2026)
- [Large Language Models for Multilingual Code Intelligence: A Survey](https://mlsanity.com/paper/2604.25960/) (2026)
- [Documentation-Guided Agentic Codebase Migration from C to Rust](https://mlsanity.com/paper/2605.14634/) (2026)

## Abstract (arXiv)

Studies show that large language models (LLMs) produce buggy code translations. One promising avenue to improve translation accuracy is through intermediate representations, which provide structured guidance for the translation process. We investigate whether LLM-based code translation can benefit from intermediate representations, specifically in the form of natural language (NL) summaries and abstract syntax trees (ASTs). Since prompt engineering greatly affects LLM performance, we consider several ways to integrate these representations, from one-shot to chain-of-thought (CoT) prompting. Using Open GPT4 8X7B and specialized StarCoder and CodeGen models on popular code translation benchmarks (CodeNet and AVATAR), we find that CoT with an intermediate NL summary performs best, with an increase of 13.8% and 6.7%, respectively, in successful translations for the best-performing model (Open GPT4 8X7B) compared to the zero-shot prompt.