Cross-Lingual Generalization and Compression: From Language-Specific to Shared Neurons
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866912409503924224 |
|---|---|
| author | Riemenschneider, Frederick Frank, Anette |
| author_facet | Riemenschneider, Frederick Frank, Anette |
| contents | Multilingual language models (MLLMs) have demonstrated remarkable abilities to transfer knowledge across languages, despite being trained without explicit cross-lingual supervision. We analyze the parameter spaces of three MLLMs to study how their representations evolve during pre-training, observing patterns consistent with compression: models initially form language-specific representations, which gradually converge into cross-lingual abstractions as training progresses. Through probing experiments, we observe a clear transition from uniform language identification capabilities across layers to more specialized layer functions. For deeper analysis, we focus on neurons that encode distinct semantic concepts. By tracing their development during pre-training, we show how they gradually align across languages. Notably, we identify specific neurons that emerge as increasingly reliable predictors for the same concepts across languages. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_01629 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Cross-Lingual Generalization and Compression: From Language-Specific to Shared Neurons Riemenschneider, Frederick Frank, Anette Computation and Language I.2.4; I.2.7 Multilingual language models (MLLMs) have demonstrated remarkable abilities to transfer knowledge across languages, despite being trained without explicit cross-lingual supervision. We analyze the parameter spaces of three MLLMs to study how their representations evolve during pre-training, observing patterns consistent with compression: models initially form language-specific representations, which gradually converge into cross-lingual abstractions as training progresses. Through probing experiments, we observe a clear transition from uniform language identification capabilities across layers to more specialized layer functions. For deeper analysis, we focus on neurons that encode distinct semantic concepts. By tracing their development during pre-training, we show how they gradually align across languages. Notably, we identify specific neurons that emerge as increasingly reliable predictors for the same concepts across languages. |
| title | Cross-Lingual Generalization and Compression: From Language-Specific to Shared Neurons |
| topic | Computation and Language I.2.4; I.2.7 |
| url | https://arxiv.org/abs/2506.01629 |