# Discovering Translation-Worthy Languages with E-Values

arXiv: https://arxiv.org/abs/2609.06593 · HTML: https://mlsanity.com/paper/2609.06593/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 14 of its 16 references that are in the map.
- First posted on arXiv 6 Sep 2026.

## Builds on (most cited first, 14 of 14)

- [RouteLLM: Learning to Route LLMs with Preference Data](https://mlsanity.com/paper/2406.18665/) (2024)
- [Game-theoretic statistics and safe anytime-valid inference](https://mlsanity.com/paper/2210.01948/) (2022)
- [SIB-200: A Simple, Inclusive, and Big Evaluation Dataset for Topic Classification in 200+ Languages and Dialects](https://mlsanity.com/paper/2309.07445/) (2023)
- [MASSIVE: A 1M-Example Multilingual Natural Language Understanding Dataset with 51 Typologically-Diverse Languages](https://mlsanity.com/paper/2204.08582/) (2022)
- [Safe Testing](https://mlsanity.com/paper/1906.07801/) (2019)
- [Revisiting Machine Translation for Cross-lingual Classification](https://mlsanity.com/paper/2305.14240/) (2023)
- [FrugalML: How to Use ML Prediction APIs More Accurately and Cheaply](https://mlsanity.com/paper/2006.07512/) (2020)
- [Confidence sequences for sampling without replacement](https://mlsanity.com/paper/2006.04347/) (2020)
- [PACE: Anytime-Valid Acceptance Tests for Self-Evolving Agents](https://mlsanity.com/paper/2606.08106/) (2026)
- [Comparing Sequential Forecasters](https://mlsanity.com/paper/2110.00115/) (2021)
- [Resolution Diagnostics for Paired LLM Evaluation](https://mlsanity.com/paper/2605.30315/) (2026)
- [RouteLMT: Learned Sample Routing for Hybrid LLM Translation Deployment](https://mlsanity.com/paper/2604.22520/) (2026)
- [No One Fits All: From Fixed Prompting to Learned Routing in Multilingual LLMs](https://mlsanity.com/paper/2604.16937/) (2026)
- [A Cost-Efficient Routing Pipeline for Multilingual Short-Text Classification Using Small Language Models](https://mlsanity.com/paper/2608.10939/) (2026)

## Abstract (arXiv)

Choosing when to translate multilingual documents is a central routing problem in text classification: translation can improve predictions for some languages while degrading others or adding unnecessary computation. Uniform translation and heuristic language tiers do not provide statistically controlled route selection. We introduce a language-level router based on paired e-processes that continuously compares direct and translation-assisted classification before freezing a routing policy. A familywise-controlled threshold of 280 bounds the probability of any false route across 14 eligible languages per dataset by 0.05. On SIB-200 and MASSIVE, the router selects translation for 4 of 15 languages and 14 of 15 locales, improving held-out accuracy over direct classification by 8.14 and 16.70 percentage points, respectively. All 28 decisions remain stable across 50 outcome-independent orderings and relative to the per-group threshold. Our results demonstrate that paired e-processes enable statistically controlled, anytime-valid, and auditable multilingual classification routing.