Skip to content

R0027/2026-03-26/Q002/S02/R01

Research R0027 — Multilingual prompt engineering challenges
Run 2026-03-26
Query Q002
Search S02
Result S02-R01

Quantifies tokenization cost per language structure

Summary

Field Value
Title The Token Tax: Systematic Bias in Multilingual Tokenization
URL https://arxiv.org/html/2509.05486v1
Date accessed 2026-03-26
Publication date 2025-09
Author(s) Jessica M. Lundin et al.
Publication arXiv preprint

Selection Decision

Included in evidence base: Yes

Rationale: Quantifies how morphological complexity translates to tokenization cost and accuracy loss.