Metraipolis › archive › dc_4ed0eb76d8d8

The More It Says, the More You Pay: A Black-Box Audit of Provider-Side Token Inflation in LLM Services — arXiv:2609.20370

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.20370
License (as archived)
CC0-1.0
Archived
2026-10-10T04:00:06.263Z
SHA-256 of the archived text
0bcc8bb3aa9ace847897de9d3a921463d910d167e525b843c3f77f9363065cee
Stable URL of this page
https://metraipolis.com/documento/dc_4ed0eb76d8d8

How to cite

The More It Says, the More You Pay: A Black-Box Audit of Provider-Side Token Inflation in LLM Services — arXiv:2609.20370 — arXiv (en), https://arxiv.org/abs/2609.20370 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_4ed0eb76d8d8 (Metraipolis archive dc_4ed0eb76d8d8, sha256 0bcc8bb3aa9ace847897de9d3a921463d910d167e525b843c3f77f9363065cee)
BibTeX
@misc{dc4ed0eb76d8d8,
  title = {The More It Says, the More You Pay: A Black-Box Audit of Provider-Side Token Inflation in LLM Services — arXiv:2609.20370},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.20370},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_4ed0eb76d8d8; sha256 0bcc8bb3aa9ace847897de9d3a921463d910d167e525b843c3f77f9363065cee},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.20370. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

In pay-per-token LLM services, the more a model says, the more users pay. Dishonest providers can covertly manipulate generation to inflate output tokens while largely preserving task utility. We define such manipulation as a Provider-Side Token Inflation Attack (PTIA) and instantiate five representative attacks at the query, prompt, representation, and model levels of the provider-controlled pipeline. Our experiments show that each attack increases mean output length to more than 10.2x the clean baseline, demonstrating PTIA's financial appeal and feasibility at multiple stages of generation. Yet auditing PTIA from black-box responses is difficult for users. Our key observation is PTIA saturation: an initial attack sharply lengthens output, but further strengthening or composition has much less effect. We trace this saturation to stopping behavior: an initial PTIA sharply lowers the end-of-sequence token probability, whereas further intervention lowers it only marginally. Building on this insight, we design a lightweight single-probe audit that applies a controlled lengthening intervention. Under PTIA, the probe induces far fewer additional tokens than under normal service. The audit requires neither a trusted local reference model nor historical clean responses, and its separately issued original and probed requests resemble ordinary traffic, making evasion difficult. Across four open-weight models, it achieves an average detection rate of 85.1% with false-positive rates below 2%. Across 15 real LLM API services, the audit flags 7 for PTIA-consistent behavior.

Tema degli esploratori: fiducia digitale e crittografia (retro-catalogo arXiv: annunci del 2026-09-18 in cs.CR).
Categorie arXiv: Cryptography and Security (cs.CR).
Pagina del lavoro: https://arxiv.org/abs/2609.20370 — 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 2026-10-10T03:26:20.885Z dalla pagina https://arxiv.org/catchup/cs.CR/2026-09-18?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).