@article{registeralwaysmattersanalysisofllm, title = {Register Always Matters: Analysis of LLM Pretraining Data Through the Lens of Language Variation}, author = {Amanda Myntti and Erik Henriksson and Veronika Laippala and Sampo Pyysalo}, year = {2025}, eprint = {2504.01542}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2504.01542v1}, }