English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

XBridge: Composing LLMs with Pretrained Encoder-Decoder Translation Models for Balanced Multilingual Capability

Forum topic · 小凯 · 2026-03-19

Summary

A Chinese tech forum post introduces XBridge (arXiv:2503.13831), a paper by Mengyu Bu and Yang Feng published March 18, 2025. Large language models show strong general intelligence but highly imbalanced multilingual performance, often failing to connect their knowledge with low-resource or unseen languages. XBridge is a compositional encoder-LLM-decoder architecture that offloads multilingual understanding and generation to an external pretrained encoder-decoder translation model, while keeping the LLM as an English-centric core for general knowledge processing. To fix representation misalignment across models, the authors introduce lightweight cross-model mapping layers and an optimal-transport-based alignment objective for fine-grained semantic consistency in multilingual generation. Experiments on four LLMs covering multilingual understanding, reasoning, summarization, and generation show XBridge outperforms strong baselines, especially on low-resource and previously unseen languages, all without retraining the LLM.

Paper Overview

Field: NLP Authors: Mengyu Bu, Yang Feng Published: 2025-03-18 arXiv: 2503.13831

Summary

Large language models (LLMs) exhibit strong general intelligence, yet their multilingual performance remains highly imbalanced. Although LLMs encode substantial cross-lingual knowledge in a unified semantic space, they often struggle to reliably interface this knowledge with low-resource or unseen languages. Fortunately, pretrained encoder-decoder translation models already possess balanced multilingual capability, suggesting a natural complement to LLMs.

Approach

The authors propose XBridge, a compositional encoder-LLM-decoder architecture that:

  • Offloads multilingual understanding and generation to external pretrained translation models
  • Preserves the LLM as an English-centric core for general knowledge processing
  • Introduces lightweight cross-model mapping layers to address representation misalignment across models
  • Uses an optimal-transport-based alignment objective to achieve fine-grained semantic consistency in multilingual generation

Results

Experiments across four LLMs on multilingual understanding, reasoning, summarization, and generation tasks show that XBridge outperforms strong baselines, particularly on low-resource and previously unseen languages — all without retraining the LLM.

---

*Auto-collected on 2026-03-19*

Tags

#llm#nlp#multilingual#machine-translation#xbridge#arxiv#optimal-transport#paper

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/177168907