aiqbench analysis
This research introduces Centroid Intervention Fusion (CIF), a scalable method that consolidates multilingual intervention projections into a single shared operator to enhance low-resource language performance without retraining models. By improving cross-lingual transfer across commonsense reasoning, inference, and translation benchmarks, the approach addresses critical scalability issues in current pairwise intervention techniques. The work represents a significant step toward more equitable and efficient large language model deployment globally.
Cross-lingual Representation Learning via Centroid Intervention Fusion
arXiv:2608.26357v1 Announce Type: new Abstract: Large language models (LLMs) exhibit uneven multilingual performance, especially when dealing with low-resource languages. Inference-time intervention offers a lightweight way to improve cross-lingual transfer by modifying the hidden states produced by the LLMs during the forward pass, without updating model parameters. However, existing cross-lingual intervention methods typically learn separate projections from source to target languages, which limits scalability and prevents knowledge sharing across languages. We propose Centroid Intervention Fusion (CIF), a projection fusion framework that consolidates multiple multilingual intervention projections into a single language-shared operator. Across multilingual commonsense reasoning, natural language inference, factual editing, and machine translation benchmarks, CIF outperforms the strongest prior pairwise intervention baseline by up to +3.378 pp on average across four model backbones, while supporting performance gains for low resource languages. The code is available at https://github.com/VRCMF/CIF.git.
Abstract:Large language models (LLMs) exhibit uneven multilingual performance, especially when dealing with low-resource languages. Inference-time intervention offers a lightweight way to improve cross-lingual transfer by modifying the hidden states produced by the LLMs during the forward pass, without updating model parameters. However, existing cross-lingual intervention methods typically learn separate projections from source to target languages, which limits scalability and prevents knowledge sharing across languages. We propose Centroid Intervention Fusion (CIF), a projection fusion framework that consolidates multiple multilingual intervention projections into a single language-shared operator. Across multilingual commonsense reasoning, natural language inference, factual editing, and machine translation benchmarks, CIF outperforms the strongest prior pairwise intervention baseline by up to +3.378 pp on average across four model backbones, while supporting performance gains for low resource languages. The code is available at this https URL.
Computation and Language (cs.CL)
arXiv:2608.26357 [cs.CL]
(or arXiv:2608.26357v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2608.26357 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
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 · 228 words
Related coverage
Through the LENS: Local Geometric Decomposition of Vision-Language Model Representations
arXiv · Aug 5, 2026
Inference-Time Steering for Cross-Lingual Factual Consistency in LLMs
arXiv · Jul 23, 2026
Language Steering for Multilingual In-Context Learning
arXiv · Oct 15, 2026
Latent-Space Intervention for Cross-Lingual Factual Consistency: Consistency Improvements without Accuracy Drops
arXiv · Sep 1, 2026