Communication Dans Un Congrès Année : 2026

Token-Efficient Change Detection in LLM APIs

Résumé

Remote change detection in LLMs is a difficult problem. Existing methods are either too expensive for deployment at scale, or require initial white-box access to model weights or grey-box access to log probabilities. We aim to achieve both low cost and strict black-box operation, observing only output tokens. Our approach hinges on specific inputs we call Border Inputs, for which there exists more than one output top token. From a statistical perspective, optimal change detection depends on the model's Jacobian and the Fisher information of the output distribution. Analyzing these quantities in low-temperature regimes shows that border inputs enable powerful change detection tests. Building on this insight, we propose the Black-Box Border Input Tracking (B3IT) scheme. Extensive in-vivo and in-vitro experiments show that border inputs are easily found for non-reasoning tested endpoints, and achieve performance on par with the best available grey-box approaches. B3IT reduces costs by 30× compared to existing methods, while operating in a strict black-box setting.

Fichier principal
Vignette du fichier
main.pdf (670.9 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Licence

Dates et versions

hal-05506096 , version 1 (11-02-2026)

Licence

Identifiants

  • HAL Id : hal-05506096 , version 1

Citer

Timothee Chauvin, Clément Lalanne, Erwan Le Merrer, Jean-Michel Loubes, François Taïani, et al.. Token-Efficient Change Detection in LLM APIs. ICML 2026 - 43rd International Conference on Machine Learning, Jul 2026, Seoul (Korea), South Korea. ⟨hal-05506096⟩
744 Consultations
467 Téléchargements

Partager

  • More