arXiv Computation and Language

Beetle: A Bilingual Model Suite for Modelling Second-Language Processing

arXiv Computation and Language
Aug 28

Double Trouble: Bilingual Pretraining Leaves Language-Conditioned Effects in Shared-Language Representations

The study compares an English-only and a bilingual decoder-only model, each 310 M parameters, trained on eight diverse languages while controlling for English exposure, compute, and document overlap. After aligning on shared English vocabulary, the authors find that token embeddings appear similar, but the deeper hidden states used for prediction diverge across models. This hidden‑state mismatch grows through middle transformer layers and persists despite controls, indicating that contextual processing differs between the models. "whyItMatters":"The findings show that embedding alignment can conceal significant internal representation differences, which is crucial for any downstream work that assumes aligned multilingual models are interchangeable."

By Anjishnu Mukherjee, Ziwei Zhu, Antonios Anastasopoulos
arXiv Computation and Language
Aug 28

Cross-Lingual Alignment Without Joint Training: Do Monolingual Language Models Converge on Universal Representations?

The study investigates whether monolingual language models, trained without joint multilingual objectives, develop cross-lingual alignment. By evaluating models such as Goldfish and independently built monolingual systems, the authors find that alignable representational geometry emerges across layers, strengthening with larger data, larger models, or closer linguistic proximity. A single Procrustes rotation on parallel sentences can map hidden states between models, and applying this rotation to a German model’s residuals swaps factual predictions to those of the donor English model, demonstrating functional transfer.

By Ej Zhou, Suchir Salhan, Catherine Arnett, Anna Korhonen
arXiv Computation and Language
Aug 27

One Form to Transfer Them All: Pretraining Multilingual Language Models Beyond Native Orthography

The study evaluates how different input representations—orthographic text, IPA transcription, and romanization—affect cross‑lingual transfer in autoregressive multilingual language models. Across three model sizes and eight languages grouped into typologically motivated pairs, romanized pretraining consistently outperforms native orthography and IPA, especially as model scale increases. Fine‑tuning a text‑pretrained model on romanized data can harm performance on languages already covered by the base model, suggesting romanization should be integrated at pretraining rather than applied later.

By Muge Zhang, Aaron Jencks, Krishna Badikela, Yulia Tsvetkov, Sachin Kumar
arXiv Computation and Language
Oct 12

Copy First, Translate Later: Interpreting Translation Dynamics in Multilingual Pretraining

Large language models show strong cross‑lingual abilities, but previous studies have examined them only at isolated training points. This work pretrains a 1.7B multilingual model on nine languages, recording checkpoints at fine granularity to track translation development. Using word‑level translation and a new dataset, the authors find that early learning is dominated by token‑level copying and surface similarities, while later stages involve more general translation mechanisms and refined copying.

By Felicia K\"orner, Maria Matveev, Florian Eichin, Gitta Kutyniok, Barbara Plank, Michael A. Hedderich
arXiv Computation and Language
Sep 28

Feeding BabyLMs Macaroni: Code-Switching Curricula Cause Cross-Lingual Convergence

The paper investigates whether training small decoder-only transformers on code‑switched text can induce cross‑lingual alignment. Using two 100‑million‑word multilingual corpora—one a mix of English, Dutch, and Chinese BabyBabelLM data, and another generated by inserting word‑ and sentence‑level code‑switching via an LLM—the authors find that code‑switched training aligns representations of parallel text, especially across different scripts, and that this alignment persists when later training on monolingual documents. A curriculum that progresses from word‑level code‑switching to sentence‑level code‑switching and finally to monolingual data yields models that outperform baselines on the BabyLM evaluation suite, demonstrating that code‑switching curriculum learning is an effective data augmentation strategy for multilingual pretraining.

By Dries Rooryck, Alex Cai, Yonatan Belinkov, David Alvarez-Melis, Kiant\'e Brantley
arXiv AI
Sep 4

One Model to Translate Them All? A Journey to Mount Doom for Multilingual Model Merging

The paper investigates weight‑space merging of independently fine‑tuned multilingual machine translation models. Experiments show that merging is more successful when models share a target language, yet it still cannot match the peak performance of language‑specific checkpoints. When target languages differ, performance drops sharply, and analysis reveals that overlapping neuron activation and incompatible upper‑layer geometries cause these failures.

By Baban Gain, Trilok Nath Singh, Asif Ekbal
arXiv Computation and Language
Oct 23

Bilingual Language Models Capture Cross-Linguistic Influence in Second-Language Reading

The study examines whether bilingual language models can reflect the influence of a reader’s first language (L1) on second-language (L2) reading. By comparing bilingual models trained on an L1 that matches the readers’ backgrounds to monolingual English models and bilingual models trained on other L1s, the authors find that matching L1 improves predictions of L2 reading times, especially in late-stage eye-tracking measures and among intermediate-proficiency readers. The effect is modulated by the overlap between L1 and L2 scripts, indicating that bilingual models capture cross-linguistic influence in L2 reading behavior.

By Abderrahmane Issam, Irene E. Winther, Stefan L. Frank
arXiv Computation and Language
Sep 30

Cross-Linguistic Effects in Bilingual Phoneme BabyLMs

arXiv:2609.37121v1 Announce Type: new Abstract: Cross-linguistic effects are a central topic in bilingual first-language acquisition. Artificial learners can help investigate L1-L2 interactions by en...

By Nikitas Theodoropoulos, Maria Lymperaiou, Giorgos Filandrianos
arXiv AI
Sep 18

Why Pretraining Fails to Share Cross-Lingual Knowledge

The paper investigates why large language models (LLMs) fail to transfer knowledge across languages. By pretraining 360M- and 7B-parameter models, the authors show that poor cross‑lingual generalization arises during pretraining and persists despite standard fixes. Using a controlled bilingual setup with identical text but disjoint token spaces, they demonstrate that token disjointness alone causes knowledge compartmentalization, and that mapping languages into a shared token space via word‑wise translation markedly improves cross‑lingual performance, recovering up to 12.6% of native‑language learning efficiency.

By Adam Gaber, Uriel Dolev, Elisabeth Fittschen, Bobby Cheng, Yuval Marton, Leshem Choshen
arXiv Computation and Language
Sep 7

Choosing the Right Language Mode at Inference Time for Multilingual Reliability

The paper investigates how multilingual large language models can be guided to reason more reliably in low- to mid-resource languages by selecting appropriate language modes during inference. Experiments with LLaMA and Qwen models show that using English context can correct errors from non‑English comprehension, but adding redundant bilingual context can cause interference. To balance this trade‑off, the authors propose Reliability‑Aware Adaptive Inference (RAAI), a training‑free test‑time framework that routes prompts based on Expected Calibration Error and gates reasoning with a mid‑layer Risk Index, achieving up to 37.7% accuracy gains and reduced calibration error on low‑resource languages.

By Ekata Mitra, Ameeta Agrawal
arXiv Computation and Language
Nov 9

Outgrowing the Curse of Multilinguality: Scaling Strategies for Efficient Language Adaptation

The paper explores how scaling pretrained language models can efficiently adapt them to new target languages, addressing the curse of multilinguality that hampers performance in medium- and low-resource settings. Through extensive ablations matched for data and FLOPs, the authors show that upscaling from English base models improves target-language performance by alleviating capacity constraints, especially for higher-resourced languages. While an initial upscaling step in English benefits low-resource languages, directly scaling up on the target language is more compute‑efficient for higher-resource scenarios, and for multilingual continued pre‑training, scaling up with target data yields the best balance of data efficiency and reduced English forgetting.

By Kevin Glocker, K\"atriin Kukk, Romina Oji, Marcel Bollmann, Marco Kuhlmann, Jenny Kunz