CLaS-Bench: A Cross-Lingual Alignment and Steering Benchmark

• URL: http://arxiv.org/abs/2601.08331v1
• Date: Tue, 13 Jan 2026 08:42:03 GMT
• Title: CLaS-Bench: A Cross-Lingual Alignment and Steering Benchmark
• Authors: Daniil Gurgurov, Yusser Al Ghussin, Tanja Baeumel, Cheng-Ting Chou, Patrick Schramowski, Marius Mosbach, Josef van Genabith, Simon Ostermann,
• Abstract summary: We introduce CLaS-Bench, a benchmark for evaluating language-forcing behavior in large language models (LLMs) across 32 languages.<n>We find that across languages simple residual-based DiffMean method consistently outperforms all other methods.
• Score: 21.574271160875046
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Understanding and controlling the behavior of large language models (LLMs) is an increasingly important topic in multilingual NLP. Beyond prompting or fine-tuning, , i.e.,~manipulating internal representations during inference, has emerged as a more efficient and interpretable technique for adapting models to a target language. Yet, no dedicated benchmarks or evaluation protocols exist to quantify the effectiveness of steering techniques. We introduce CLaS-Bench, a lightweight parallel-question benchmark for evaluating language-forcing behavior in LLMs across 32 languages, enabling systematic evaluation of multilingual steering methods. We evaluate a broad array of steering techniques, including residual-stream DiffMean interventions, probe-derived directions, language-specific neurons, PCA/LDA vectors, Sparse Autoencoders, and prompting baselines. Steering performance is measured along two axes: language control and semantic relevance, combined into a single harmonic-mean steering score. We find that across languages simple residual-based DiffMean method consistently outperforms all other methods. Moreover, a layer-wise analysis reveals that language-specific structure emerges predominantly in later layers and steering directions cluster based on language family. CLaS-Bench is the first standardized benchmark for multilingual steering, enabling both rigorous scientific analysis of language representations and practical evaluation of steering as a low-cost adaptation alternative.

Related papers

• What Language is This? Ask Your Tokenizer [32.28976119949841]
  Language Identification (LID) is an important component of many multilingual natural language processing pipelines.<n>We introduce UniLID, a simple and efficient LID method based on the UnigramLM tokenization algorithm.<n>Our formulation is data- and compute-efficient, supports incremental addition of new languages without retraining existing models.
  arXiv  Detail & Related papers  (2026-02-19T18:58:39Z)
• Language Steering for Multilingual In-Context Learning [10.932074928744568]
  Large language models' performance on non-English languages remains substantially inferior to English.<n>We propose language vectors -- a training-free language steering approach.<n>We show consistent improvements on multilingual in-context learning over baselines across all tasks and languages tested.
  arXiv  Detail & Related papers  (2026-02-02T16:52:09Z)
• LangGPS: Language Separability Guided Data Pre-Selection for Joint Multilingual Instruction Tuning [49.22807995935406]
  Joint multilingual instruction tuning is a widely adopted approach to improve the multilingual instruction-following ability and downstream performance of large language models (LLMs)<n>Existing selection methods, often based on features like text quality, diversity, or task relevance, typically overlook the intrinsic linguistic structure of multilingual data.<n>We propose LangGPS, a lightweight two-stage pre-selection framework guided by language separability which quantifies how well samples in different languages can be distinguished in the model's representation space.
  arXiv  Detail & Related papers  (2025-11-13T12:02:32Z)
• Language steering in latent space to mitigate unintended code-switching [1.1330938617817454]
  Large Language Models (LLMs) often exhibit unintended code-switching, reducing reliability in downstream tasks.<n>We propose latent-space language steering, a lightweight inference-time method that identifies language directions via PCA on parallel translations.<n>Our approach mitigates code-switching while preserving semantics with negligible computational overhead.
  arXiv  Detail & Related papers  (2025-10-11T19:49:38Z)
• Improving Multilingual Language Models by Aligning Representations through Steering [10.159957091670883]
  This paper investigates how Large Language Models (LLMs) represent non-English tokens.<n>We propose a lightweight intervention method using representation steering, where a learned vector is added to the residual stream at a single model layer to enhance multilingual performance.
  arXiv  Detail & Related papers  (2025-05-19T00:14:43Z)
• Investigating and Scaling up Code-Switching for Multilingual Language Model Pre-Training [58.696660064190475]
  We find that the existence of code-switching, alternating between different languages within a context, is key to multilingual capabilities.<n>To better explore the power of code-switching for language alignment during pre-training, we investigate the strategy of synthetic code-switching.
  arXiv  Detail & Related papers  (2025-04-02T15:09:58Z)
• Assessing Code Generation with Intermediate Languages [6.999311675957218]
  This study explores the utilization of intermediate languages, including various programming languages, natural language solutions, and pseudo-code. Our findings reveal that intermediate languages generally exhibit greater efficacy in larger models that have not yet achieved state-of-the-art performance.
  arXiv  Detail & Related papers  (2024-07-07T15:35:41Z)
• Analyzing and Adapting Large Language Models for Few-Shot Multilingual NLU: Are We There Yet? [82.02076369811402]
  Supervised fine-tuning (SFT), supervised instruction tuning (SIT) and in-context learning (ICL) are three alternative, de facto standard approaches to few-shot learning. We present an extensive and systematic comparison of the three approaches, testing them on 6 high- and low-resource languages, three different NLU tasks, and a myriad of language and domain setups. Our observations show that supervised instruction tuning has the best trade-off between performance and resource requirements.
  arXiv  Detail & Related papers  (2024-03-04T10:48:13Z)
• Soft Language Clustering for Multilingual Model Pre-training [57.18058739931463]
  We propose XLM-P, which contextually retrieves prompts as flexible guidance for encoding instances conditionally. Our XLM-P enables (1) lightweight modeling of language-invariant and language-specific knowledge across languages, and (2) easy integration with other multilingual pre-training methods.
  arXiv  Detail & Related papers  (2023-06-13T08:08:08Z)
• CUGE: A Chinese Language Understanding and Generation Evaluation Benchmark [144.05723617401674]
  General-purpose language intelligence evaluation has been a longstanding goal for natural language processing. We argue that for general-purpose language intelligence evaluation, the benchmark itself needs to be comprehensive and systematic. We propose CUGE, a Chinese Language Understanding and Generation Evaluation benchmark with the following features.
  arXiv  Detail & Related papers  (2021-12-27T11:08:58Z)
• Multilingual Chart-based Constituency Parse Extraction from Pre-trained Language Models [21.2879567125422]
  We propose a novel method for extracting complete (binary) parses from pre-trained language models. By applying our method on multilingual PLMs, it becomes possible to induce non-trivial parses for sentences from nine languages.
  arXiv  Detail & Related papers  (2020-04-08T05:42:26Z)
• XTREME: A Massively Multilingual Multi-task Benchmark for Evaluating Cross-lingual Generalization [128.37244072182506]
  Cross-lingual TRansfer Evaluation of Multilinguals XTREME is a benchmark for evaluating the cross-lingual generalization capabilities of multilingual representations across 40 languages and 9 tasks. We demonstrate that while models tested on English reach human performance on many tasks, there is still a sizable gap in the performance of cross-lingually transferred models.
  arXiv  Detail & Related papers  (2020-03-24T19:09:37Z)

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.