Metraipolis › archive › dc_ebd6b91c2bd6

Calibrating Post-Training Feature Shifts for LLM Data Contamination Detection — arXiv:2608.10462

Source: arXiv (en)
Original: https://arxiv.org/abs/2608.10462
License (as archived): CC0-1.0
Archived: 2026-10-19T06:14:12.060Z
SHA-256 of the archived text: cbe3aa715a3bc9b89feba6586f11af39642d125b09e68bc90206dcfc79555ffc
Stable URL of this page: https://metraipolis.com/documento/dc_ebd6b91c2bd6

How to cite

Calibrating Post-Training Feature Shifts for LLM Data Contamination Detection — arXiv:2608.10462 — arXiv (en), https://arxiv.org/abs/2608.10462 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ebd6b91c2bd6 (Metraipolis archive dc_ebd6b91c2bd6, sha256 cbe3aa715a3bc9b89feba6586f11af39642d125b09e68bc90206dcfc79555ffc)

BibTeX

@misc{dcebd6b91c2bd6,
  title = {Calibrating Post-Training Feature Shifts for LLM Data Contamination Detection — arXiv:2608.10462},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.10462},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ebd6b91c2bd6; sha256 cbe3aa715a3bc9b89feba6586f11af39642d125b09e68bc90206dcfc79555ffc},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.10462. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large language models (LLMs) are trained on massive and largely undisclosed corpora that may contain copyrighted or privacy-sensitive content. Data contamination detection (DCD) therefore aims to determine whether a given text is a member of the pre-training corpus of a target LLM. Recent state-of-the-art DCD methods follow a feature-based paradigm that derives membership features from the input text and the corresponding model output. However, most modern LLMs undergo post-training, such as instruction tuning, preference optimization, and reasoning-oriented training, which can alter model outputs and shift the corresponding membership features, thereby reducing the separability between members and non-members. To address this problem, we propose CalibDCD, a broadly applicable calibration framework for feature-based DCD methods, comprising (1) Multi-View Shift Detection, which identifies recurring feature shifts associated with post-training, and (2) Bounded Feature Correction, which selectively mitigates their influence on membership prediction. Specifically, Multi-View Shift Detection evaluates controlled prompt variants on known non-member texts and consolidates the most informative views to identify recurring feature shifts. Bounded Feature Correction selectively adjusts feature components aligned with the detected shifts and controls the correction extent to preserve useful detection information. Experiments show that CalibDCD consistently improves existing feature-based detectors, with gains of up to 7.0% in AUC and 15.0% in TPR@5%FPR.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-12 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2608.10462 — 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 2026-10-19T06:08:50.316Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-12?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).