Metraipolis › archive › dc_6ffdd6c9ff65

Improving LLM-Based SSH Honeypots Through Prompting and Fine-Tuning — arXiv:2608.18686

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.18686
License (as archived)
CC0-1.0
Archived
2026-10-18T13:51:51.568Z
SHA-256 of the archived text
4480b47f572b5ab14155dec9cd76c01b57b79f9108ec6fac3de82ad8086b0460
Stable URL of this page
https://metraipolis.com/documento/dc_6ffdd6c9ff65

How to cite

Improving LLM-Based SSH Honeypots Through Prompting and Fine-Tuning — arXiv:2608.18686 — arXiv (en), https://arxiv.org/abs/2608.18686 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_6ffdd6c9ff65 (Metraipolis archive dc_6ffdd6c9ff65, sha256 4480b47f572b5ab14155dec9cd76c01b57b79f9108ec6fac3de82ad8086b0460)
BibTeX
@misc{dc6ffdd6c9ff65,
  title = {Improving LLM-Based SSH Honeypots Through Prompting and Fine-Tuning — arXiv:2608.18686},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.18686},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_6ffdd6c9ff65; sha256 4480b47f572b5ab14155dec9cd76c01b57b79f9108ec6fac3de82ad8086b0460},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.18686. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

LLM-based SSH honeypots often use closed cloud LLMs because they give strong shell realism, but cloud models create deployment problems. These include no stable versioning, provider-side changes, attacker-driven cost, and model decommissioning. Local open-weight models avoid these problems, but they usually perform worse and make mistakes that reveal the honeypot. These mistakes include malformed outputs, command echoing, inconsistent filesystem state, and AI-style artifacts. This paper studies how to improve and evaluate the shell emulation accuracy of local LLM-based SSH honeypots using prompt design and supervised fine-tuning. We fine-tune and evaluate eight models in total: the original fine-tuned GPT-3.5 model used in shelLM and seven open-weight local models, each compared to its base model. We also test how prompt structure transfers across model families. Using 34 automated unit tests that measure shell emulation accuracy in single-session and fresh-session settings, we find that prompt design has a large effect and that fine-tuning depends on dataset coverage. Fine-tuning on the original 112-conversation dataset does not improve aggregate pass rate, while an expanded dataset built from honeypot logs produces clearly stronger local models. Taken together, the results suggest that prompting and fine-tuning can each improve local LLM honeypots on their own, but their effects do not combine straightforwardly, since strong rule-based prompting and supervised adaptation can also conflict by addressing overlapping shell-behavior constraints.

Tema degli esploratori: fiducia digitale e crittografia (retro-catalogo arXiv: annunci del 2026-08-20 in cs.CR).
Categorie arXiv: Cryptography and Security (cs.CR).
Pagina del lavoro: https://arxiv.org/abs/2608.18686 — 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 2026-10-18T13:49:54.231Z dalla pagina https://arxiv.org/catchup/cs.CR/2026-08-20?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).