Skip to content
Preprint

Cross-lingual Representation Learning via Centroid Intervention Fusion

Aug 2026 · 0 citations · 41 references
Computer Science

TL;DR

Centroid Intervention Fusion is proposed, a projection fusion framework that consolidates multiple multilingual intervention projections into a single language-shared operator and outperforms the strongest prior pairwise intervention baseline by up to +3.3% across four model backbones.

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.

View source

Similar papers

Preprint Aug 2026

Native Multilingual Chain-of-Thought Reasoning in Low-Resource Southeast Asian Languages

The Onramp-Sequence Cross-Distillation (OSCD) is introduced, a post-training algorithm that projects high-resource reasoning trajectories into low-resource vocabulary subspaces during generative training rollouts via an integrated translator agentic loop, ensuring the stable and efficient translation of dynamically generated reference samples for fine-tuning.

Sean Gip Lim, William-Chandra Tjhi, Hai Leong Chieu · 0 citations
Preprint Jul 2026

Inference-Time Steering for Cross-Lingual Factual Consistency in LLMs

It is suggested that cross-lingual inconsistency is at least partly a selection problem, and that simple contextual interventions may outperform more invasive methods for robust, transferable alignment.

Alexander Manev · 0 citations
Preprint Jul 2026

Unified Gradient Projection: Language-Balanced Continual Learning for Multilingual Low-Resource ASR

Unified Gradient Projection is proposed, which constrains parameter updates using reference gradients from language-balanced replay in a unified projection space and shows that combining gradient-level projection with data-level replay yields complementary gains in stability and plasticity.

Ziang Ren, Guodong Lin, Yuchen Ai et al. · 0 citations
Preprint Aug 2026

When the API Speaks the Wrong Language: Revisiting Post-Training for Multilingual Tool Use

It is found that, in this benchmark, supervised fine-tuning (SFT) provides a strong baseline, substantially improving argument language consistency and end-to-end function call accuracy and, under consistent model selection, SFT achieves performance comparable to, and sometimes exceeding more complex reinforcement learning (RL) approaches.

Siddharth Chauhan, Thomas Butler, Abhishek Singhania et al. · 0 citations
Preprint Aug 2026

Strengthening Target-Language Features: SAE-Based Steering for Multilingual Inference

Multilingual large language models exhibit substantial performance differences across languages, while existing adaptation methods often require parameter updates and considerable multilingual training data. We propose an inference-time multilingual steering method that uses pretrained sparse autoencoders to identify and strengthen target-language-related features. Using multilingual parallel sentences, we compare SAE activations across languages and select a small number of layer-specific features associated with each target language. These features are decoded into steering signals and injected into the model's hidden states without additional training. Experiments with Gemma-3-12B-it show average accuracy improvements of 10.9 percentage points on XCOPA, 5.3 points on XNLI, and 1.9 points on MGSM.

Hongsheng Wang, Philipp Koehn · 0 citations
Open access Jul 2026

Automated Multilingual Translator Using Neural Translation

The results indicate that a moderately sized, shared self-attention architecture can deliver production-quality multilin-gual translation within the resource constraints of an academic de-ployment, while surfacing clear directions – low-resource language coverage, domain adaptation, and speech-based extension – for con-tinued development.

Darshan Gowda D H and Dr. Kruti R · 0 citations