Pré-Publication, Document De Travail Année : 2025

Date Fragments: A Hidden Bottleneck of Tokenisation for Temporal Reasoning

Résumé

Modern BPE tokenisers often split calendar dates into meaningless fragments, e.g., "20250312" → "202", "503", "12", inflating token counts and obscuring the inherent structure needed for robust temporal reasoning. In this work, we (1) introduce a simple yet interpretable metric, termed date fragmentation ratio, that measures how faithfully a tokeniser preserves multi-digit date components;

(2) release DATEAUGBENCH, a suite of 6500 examples spanning three temporal reasoning tasks: context-based date resolution, formatinvariance puzzles, and date arithmetic across historical, contemporary, and future time periods; and (3) through layer-wise probing and causal attention-hop analyses, uncover an emergent date-abstraction mechanism whereby large language models stitch together the fragments of month, day, and year components for temporal reasoning. Our experiments show that excessive fragmentation correlates with accuracy drops of up to 10 points on uncommon dates like historical and futuristic dates. Further, we find that the larger the model, the faster the emergent date abstraction heals date fragments. Lastly, we observe a reasoning path that LLMs follow to assemble date fragments, typically differing from human interpretation (year → month → day). Our datasets and code are made publicly available here

Fichier principal
Vignette du fichier
2505.16088v2.pdf (1.5 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Licence

Dates et versions

hal-05226368 , version 1 (27-08-2025)
hal-05226368 , version 2 (27-11-2025)

Licence

Identifiants

Citer

Gagan Bhatia, Maxime Peyrard, Wei Zhao. Date Fragments: A Hidden Bottleneck of Tokenisation for Temporal Reasoning. 2025. ⟨hal-05226368v1⟩
202 Consultations
475 Téléchargements

Altmetric

Partager

  • More