Benchmarking Concept-Spilling Across Languages in LLMs

• URL: http://arxiv.org/abs/2601.12549v1
• Date: Sun, 18 Jan 2026 19:28:26 GMT
• Title: Benchmarking Concept-Spilling Across Languages in LLMs
• Authors: Ilia Badanin, Daniil Dzenhaliou, Imanol Schlag,
• Abstract summary: Large Language Models (LLMs) exhibit remarkable cross-lingual abilities, yet often exhibit a systematic bias toward representations from other languages.<n>This paper presents a novel comparative framework for evaluating multilingual semantic robustness by measuring how models handle polysemous words across languages.
• Score: 7.577675422356702
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Multilingual Large Language Models (LLMs) exhibit remarkable cross-lingual abilities, yet often exhibit a systematic bias toward the representations from other languages, resulting in semantic interference when generating content in non-English languages$-$a phenomenon we define as language spilling. This paper presents a novel comparative framework for evaluating multilingual semantic robustness by systematically measuring how models handle polysemous words across languages. Our methodology provides a relative measure of model performance: when required to generate exactly five meanings, both strong and weak models may resort to meanings from dominant languages, but semantically stronger models do so later in the generation sequence, producing more true meanings from the target language before failing, while weaker models resort to dominant-language meanings earlier in the sequence. We evaluate a diverse set of open and closed multilingual LLMs using a structured meaning generation task across nine languages, employing a carefully curated benchmark of 100 high-polysemy English words. Our findings reveal significant variation in semantic robustness across both models and languages, providing a principled ranking system for model comparison without requiring definitive causal attribution of error sources. We contribute both a scalable comparative benchmark for multilingual semantic evaluation and a rigorous validation pipeline$-$critical tools for developing more linguistically balanced AI systems.

Related papers

• Parallel Universes, Parallel Languages: A Comprehensive Study on LLM-based Multilingual Counterfactual Example Generation [49.2073409243885]
  Large language models (LLMs) excel at generating English counterfactuals and demonstrate multilingual proficiency.<n>We conduct automatic evaluations on both directly generated counterfactuals in the target languages and those derived via English translation across six languages.<n>We identify and categorize four main types of errors that consistently appear in the generated counterfactuals across languages.
  arXiv  Detail & Related papers  (2026-01-01T08:53:49Z)
• Evaluating Large Language Models on Multiword Expressions in Multilingual and Code-Switched Contexts [2.519319150166215]
  This study evaluates how state-of-the-art language models process the ambiguity of potentially idiomatic multiword expressions.<n>We find that large language models, despite their strengths, struggle with nuanced language.
  arXiv  Detail & Related papers  (2025-04-10T16:39:28Z)
• AdaMCoT: Rethinking Cross-Lingual Factual Reasoning through Adaptive Multilingual Chain-of-Thought [40.16140566668239]
  We introduce AdaMCOT, a framework that enhances multilingual factual reasoning.<n>AdaMCOT dynamically routing thought processes in intermediary "thinking languages" before generating target-language responses.<n>Our evaluation demonstrates substantial improvements in both factual reasoning quality and cross-lingual consistency.
  arXiv  Detail & Related papers  (2025-01-27T15:48:57Z)
• Multilingual Conceptual Coverage in Text-to-Image Models [98.80343331645626]
  "Conceptual Coverage Across Languages" (CoCo-CroLa) is a technique for benchmarking the degree to which any generative text-to-image system provides multilingual parity to its training language in terms of tangible nouns. For each model we can assess "conceptual coverage" of a given target language relative to a source language by comparing the population of images generated for a series of tangible nouns in the source language to the population of images generated for each noun under translation in the target language.
  arXiv  Detail & Related papers  (2023-06-02T17:59:09Z)
• HiJoNLP at SemEval-2022 Task 2: Detecting Idiomaticity of Multiword Expressions using Multilingual Pretrained Language Models [0.6091702876917281]
  This paper describes an approach to detect idiomaticity only from the contextualized representation of a MWE over multilingual pretrained language models. Our experiments find that larger models are usually more effective in idiomaticity detection. However, using a higher layer of the model may not guarantee a better performance.
  arXiv  Detail & Related papers  (2022-05-27T01:55:59Z)
• Language Models are Few-shot Multilingual Learners [66.11011385895195]
  We evaluate the multilingual skills of the GPT and T5 models in conducting multi-class classification on non-English languages. We show that, given a few English examples as context, pre-trained language models can predict not only English test samples but also non-English ones.
  arXiv  Detail & Related papers  (2021-09-16T03:08:22Z)
• Discovering Representation Sprachbund For Multilingual Pre-Training [139.05668687865688]
  We generate language representation from multilingual pre-trained models and conduct linguistic analysis. We cluster all the target languages into multiple groups and name each group as a representation sprachbund. Experiments are conducted on cross-lingual benchmarks and significant improvements are achieved compared to strong baselines.
  arXiv  Detail & Related papers  (2021-09-01T09:32:06Z)
• Specializing Multilingual Language Models: An Empirical Study [50.7526245872855]
  Contextualized word representations from pretrained multilingual language models have become the de facto standard for addressing natural language tasks. For languages rarely or never seen by these models, directly using such models often results in suboptimal representation or use of data.
  arXiv  Detail & Related papers  (2021-06-16T18:13:55Z)
• AM2iCo: Evaluating Word Meaning in Context across Low-ResourceLanguages with Adversarial Examples [51.048234591165155]
  We present AM2iCo, Adversarial and Multilingual Meaning in Context. It aims to faithfully assess the ability of state-of-the-art (SotA) representation models to understand the identity of word meaning in cross-lingual contexts. Results reveal that current SotA pretrained encoders substantially lag behind human performance.
  arXiv  Detail & Related papers  (2021-04-17T20:23:45Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.