Metraipolis › archive › dc_8460fdce6397

Stage-Supervised Latent Reasoning for Single-Shot JavaScript Deobfuscation — arXiv:2609.27058

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.27058
License (as archived)
CC0-1.0
Archived
2026-10-08T11:06:23.698Z
SHA-256 of the archived text
015e4a48d697006ead52de93f08e0392f0362ac3f130565e93c4ee57a6d5ffc8
Stable URL of this page
https://metraipolis.com/documento/dc_8460fdce6397

How to cite

Stage-Supervised Latent Reasoning for Single-Shot JavaScript Deobfuscation — arXiv:2609.27058 — arXiv (en), https://arxiv.org/abs/2609.27058 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_8460fdce6397 (Metraipolis archive dc_8460fdce6397, sha256 015e4a48d697006ead52de93f08e0392f0362ac3f130565e93c4ee57a6d5ffc8)
BibTeX
@misc{dc8460fdce6397,
  title = {Stage-Supervised Latent Reasoning for Single-Shot JavaScript Deobfuscation — arXiv:2609.27058},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.27058},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_8460fdce6397; sha256 015e4a48d697006ead52de93f08e0392f0362ac3f130565e93c4ee57a6d5ffc8},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.27058. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

JavaScript obfuscation is widely used to protect code, but it also makes program analysis and security review substantially harder. Existing LLM-based deobfuscation methods usually treat the task as one-step translation, ignoring the staged structure of practical deobfuscation pipelines. This WIP paper proposes a stage-aware latent reasoning framework that converts intermediate outputs from a deterministic deobfuscation tool into supervision for Coconut-based training. The model learns from multi-stage rewrites during training but generates the final cleaned program in a single shot at inference time. Preliminary results on JsDeObsBench show that the Coconut-based model improves syntactic validity to 50%, compared with 15% for direct fine-tuning and 25% for a zero-shot baseline, and reaches 80% semantic correctness among valid outputs, indicating better semantic faithfulness than either comparison model in deobfuscation.

Tema degli esploratori: calcolo e infrastrutture (retro-catalogo arXiv: annunci del 2026-09-24 in cs.SE).
Categorie arXiv: Software Engineering (cs.SE).
Pagina del lavoro: https://arxiv.org/abs/2609.27058 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-08T10:13:35.106Z dalla pagina https://arxiv.org/catchup/cs.SE/2026-09-24?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).