aiqbench analysis
This research introduces Cross-Lingual Activation Steering (CLAS), a training-free method that improves multilingual model performance by selectively modulating neuron activations during inference. The study reveals that effective cross-lingual transfer relies on functional divergence between language clusters rather than strict alignment, unlocking latent capacity in existing weights without retraining. This approach offers a practical pathway to reduce performance gaps for non-dominant languages while preserving high-resource language capabilities.
Cross-Lingual Activation Steering for Multilingual Language Models
arXiv:2601.16390v2 Announce Type: replace Abstract: Large language models exhibit strong multilingual capabilities, yet significant performance gaps persist between dominant and non-dominant languages. Prior work attributes this gap to imbalances between shared and language-specific neurons in multilingual representations. We propose Cross-Lingual Activation Steering (CLAS), a training-free inference-time intervention that selectively modulates neuron activations. We evaluate CLAS on classification and generation benchmarks, achieving average improvements of 2.3% (Acc.) and 3.4% (F1) respectively, while maintaining high-resource language performance. We discover that effective transfer operates through functional divergence rather than strict alignment; performance gains correlate with increased language cluster separation. Our results demonstrate that targeted activation steering can unlock latent multilingual capacity in existing models without modification to model weights.
Abstract:Large language models exhibit strong multilingual capabilities, yet significant performance gaps persist between dominant and non-dominant languages. Prior work attributes this gap to imbalances between shared and language-specific neurons in multilingual representations. We propose Cross-Lingual Activation Steering (CLAS), a training-free inference-time intervention that selectively modulates neuron activations. We evaluate CLAS on classification and generation benchmarks, achieving average improvements of 2.3% (Acc.) and 3.4% (F1) respectively, while maintaining high-resource language performance. We discover that effective transfer operates through functional divergence rather than strict alignment; performance gains correlate with increased language cluster separation. Our results demonstrate that targeted activation steering can unlock latent multilingual capacity in existing models without modification to model weights.
Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
arXiv:2601.16390 [cs.CL]
(or arXiv:2601.16390v2 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2601.16390 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 207 words
Related coverage
Inference-Time Steering for Cross-Lingual Factual Consistency in LLMs
arXiv · Jul 23, 2026
Cross-Lingual Steering for Figurative Language Generation
arXiv · Oct 14, 2026
Language Steering for Multilingual In-Context Learning
arXiv · Oct 15, 2026
BabelSteering: Multilingual Safety Alignment via English Steering Vectors
arXiv · Aug 18, 2026