@article{anempiricalstudyonnoisydataandllmpretrai, title = {An Empirical Study on Noisy Data and LLM Pretraining Loss Divergence}, author = {Qizhen Zhang and Ankush Garg and Jakob Foerster and Niladri Chatterji and Kshitiz Malik and Mike Lewis}, year = {2026}, eprint = {2602.02400}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2602.02400}, }