Metraipolis › archive › dc_46980317aeef

Team DArgk at the 2026 ELOQUENT lab for evaluating generative language model quality: Residuals of Humanity: AI Detection Evasion via GRPO Fine-Tuning — arXiv:2609.22221

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.22221
License (as archived)
CC0-1.0
Archived
2026-10-09T06:49:19.644Z
SHA-256 of the archived text
70b1b3deb17c6f7ab6ba034822a4c76b4b51410d1d87ad6a60f317bf4269553a
Stable URL of this page
https://metraipolis.com/documento/dc_46980317aeef

How to cite

Team DArgk at the 2026 ELOQUENT lab for evaluating generative language model quality: Residuals of Humanity: AI Detection Evasion via GRPO Fine-Tuning — arXiv:2609.22221 — arXiv (en), https://arxiv.org/abs/2609.22221 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_46980317aeef (Metraipolis archive dc_46980317aeef, sha256 70b1b3deb17c6f7ab6ba034822a4c76b4b51410d1d87ad6a60f317bf4269553a)
BibTeX
@misc{dc46980317aeef,
  title = {Team DArgk at the 2026 ELOQUENT lab for evaluating generative language model quality: Residuals of Humanity: AI Detection Evasion via GRPO Fine-Tuning — arXiv:2609.22221},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.22221},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_46980317aeef; sha256 70b1b3deb17c6f7ab6ba034822a4c76b4b51410d1d87ad6a60f317bf4269553a},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.22221. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large language models (LLMs) can generate fluent and coherent text that is increasingly difficult to distinguish from human writing, motivating the development of automatic AI-generated text detectors. However, the robustness of such detectors under adversarial generation remains uncertain. This paper presents SHADE (Stochastic Human-like generation via Adversarial Detector Evasion), a reinforcement learning framework that formulates detector evasion as a policy optimization problem. Instead of applying post-hoc perturbations or prompting-based rewriting, SHADE fine-tunes an instruction-tuned LLaMA model with Group Relative Policy Optimization (GRPO), using feedback from a surrogate detector based on the PAN 2025 mdok system. Our experiments show that full fine-tuning with a small KL regularization penalty achieves $98.5\%$ surrogate evasion, compared to $1.5\%$ for the base model, while LoRA-based adaptation is substantially less effective under regularization. Linguistic analysis reveals that successful evasion is associated with shorter, simpler, and less lexically diverse outputs, suggesting that high detector evasion does not necessarily correspond to more human-like writing. In the official Voight-Kampff competition setting, our submissions ranked sixth and seventh, indicating that optimization against a single surrogate detector only partially transfers to unseen evaluation classifiers. These results highlight both the potential and limitations of reinforcement learning for adversarial AI-text generation and motivate more robust, multi-detector evaluation protocols for AI-generated text detection.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-22 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.22221 — 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 2026-10-09T06:48:44.898Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-22?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).