# Binary Decompilation LLM with Feedback-Driven Multi-Turn Refinement

> AutoDecompiler trains function-level decompilation models to revise assembly- or pseudocode-derived outputs using compilation, execution and I/O feedback. Its reinforcement-learning design combines diagnostic feedback with rewards for valid code, tested behavior, syntax and data-flow structure.

- Authors: Peipei Liu; Jian Sun; Mingzhe Xing; Yicheng Zeng; Zhaoteng Yan; Lixiao Zhang; Li Chen; Dan Li
- Venue: arXiv
- Status: Preprint
- Publication / posting date: June 15, 2026
- Date source: First arXiv submission on 2026-06-15; the reviewed v2 revision is 2026-10-03, not a separate publication.
- First arXiv submission: June 15, 2026
- arXiv: 2606.16162
- Reviewed arXiv version: v2
- Sources reviewed: 2026-10-20
- Added to the catalog: 2026-10-20
- Original paper: <https://arxiv.org/abs/2606.16162>
- PDF: <https://arxiv.org/pdf/2606.16162v2>
- Code: <https://zenodo.org/records/23120286>
- Project: <https://huggingface.co/AutoDecompiler>
- BibTeX: <https://awesomerea.org/papers/citations/paper-autodecompiler.bib>
- Catalog page: <https://awesomerea.org/papers/paper-autodecompiler/>
- Record JSON: <https://awesomerea.org/papers/paper-autodecompiler/record.json>
- Chinese version: [简体中文](https://awesomerea.org/zh/papers/paper-autodecompiler/index.md)

## Scope and limitations

Evaluation uses C adaptations of HumanEval and MBPP with validation feedback; finite tests do not establish general equivalence. HumanEval also informs reward-weight selection, and base-model pretraining exposure cannot be ruled out. Refinement is context-limited, whole-program dependencies are not recovered, and unseen toolchains or protected binaries remain limitations.

## Why it fits this collection

Adds reinforcement-learning-trained use of program feedback, with separately recorded first submission and latest revision.

## Publication and review notes

First submitted 2026-06-15; reviewed v2 is dated 2026-10-03. The arXiv record uses the shorter catalog title, while the v2 PDF/HTML use a longer title beginning “AutoDecompiler: Reinforcement-Learning-Trained”; eight authors match. Zenodo version 23120286 lists code/data archives under CC BY 4.0. The inspected Hugging Face model has no license metadata.

## Primary sources

1. <https://arxiv.org/abs/2606.16162v2>
2. <https://arxiv.org/html/2606.16162v2>
3. <https://arxiv.org/abs/2606.16162v1>
4. <https://arxiv.org/pdf/2606.16162v2>
5. <https://zenodo.org/api/records/23120286>
6. <https://huggingface.co/api/models/AutoDecompiler/AutoDecompiler-1.3B-e2e>

## About this entry

Reviewed against the paper and its primary sources. Cite the original authors using the linked paper or BibTeX.
Original editorial content: MIT. Linked projects, papers and assets retain their own rights and licenses.