English

BERT-JEPA: Reorganizing CLS Embeddings for Language-Invariant Semantics

Computation and Language 2026-01-05 v1 Artificial Intelligence Machine Learning

Abstract

Joint Embedding Predictive Architectures (JEPA) are a novel self supervised training technique that have shown recent promise across domains. We introduce BERT-JEPA (BEPA), a training paradigm that adds a JEPA training objective to BERT-style models, working to combat a collapsed [CLS] embedding space and turning it into a language-agnostic space. This new structure leads to increased performance across multilingual benchmarks.

Cite

@article{arxiv.2601.00366,
  title  = {BERT-JEPA: Reorganizing CLS Embeddings for Language-Invariant Semantics},
  author = {Taj Gillin and Adam Lalani and Kenneth Zhang and Marcel Mateos Salles},
  journal= {arXiv preprint arXiv:2601.00366},
  year   = {2026}
}

Comments

16 pages, 10 figures, 10 tables

R2 v1 2026-07-01T08:47:52.297Z