Is Multilingual LLM Watermarking Truly Multilingual? Scaling Robustness to 100+ Languages via Back-Translation

Domaine:

natural language processing

Type de record:

papermodel

Créateur:

MohGub

Hôte:

Multilingual watermarking aims to make large language model (LLM) outputs traceable across languages, yet current methods still fall short. Despite claims of cross-lingual robustness, they are evaluated only on high-resource languages. We show that existing multilingual watermarking methods are not truly multilingual: they fail to remain robust under translation attacks in medium- and low-resource languages. We trace this failure to semantic clustering, which fails when the tokenizer vocabulary contains too few full-word tokens for a given language. To address this, we introduce STEAM, a detection method that uses Bayesian optimisation to search among 133 candidate languages for the back-translation that best recovers the watermark strength. It is compatible with any watermarking method, robust across different tokenizers and languages, non-invasive, and easily extendable to new languages. With average gains of +0.23 AUC and +37% TPR@1%, STEAM provides a scalable approach toward fairer watermarking across the diversity of languages.

Visit

arxiv.org

Similaires

XLLaMA2: Scaling Linguistic Horizons of LLM by Enhancing Translation Capabilities Beyond 100 LanguagesMultilingual Projection for Parsing Truly Low-Resource LanguagesSMaLL-100: Introducing Shallow Multilingual Machine Translation Model for Low-Resource LanguagesSynthetic Data Diversity vs. Back-Translation for Multilingual NER in Low-Resource LanguagesTeaching LLMs to Abstain across Languages via Multilingual FeedbackGlot500: Scaling Multilingual Corpora and Language Models to 500 Languages

XLLaMA2: Scaling Linguistic Horizons of LLM by Enhancing Translation Capabilities Beyond 100 Languages

Large Language Models (LLMs) demonstrate remarkable translation capabilities in high-resource langua

Multilingual Projection for Parsing Truly Low-Resource Languages

We propose a novel approach to cross-lingual part-of-speech tagging and dependency parsing for truly

SMaLL-100: Introducing Shallow Multilingual Machine Translation Model for Low-Resource Languages

In recent years, multilingual machine translation models have achieved promising performance on low-resource language pairs by sharing information between similar languages, thus enabling zero-shot translation. To overcome the "curse of multilinguality", these mode

Synthetic Data Diversity vs. Back-Translation for Multilingual NER in Low-Resource Languages

Named Entity Recognition(NER) for low-resource languages aims to produce robust systems for language

Teaching LLMs to Abstain across Languages via Multilingual Feedback

Multilingual LLMs often have knowledge disparities across languages, with larger gaps in under-resou

Glot500: Scaling Multilingual Corpora and Language Models to 500 Languages

The NLP community has mainly focused on scaling Large Language Models (LLMs) vertically, i.e., makin