HPLT 3.0 releases the largest open multilingual LLM pre-training dataset at 30 trillion tokens for nearly 200 languages, with quality annotations, evaluation benchmarks for nine languages, and pre-trained models.
(2025), we calculate descriptive statisticsforHPLT3.0usingtheHPLTAnalyticstool 6 and compare HPLT 3.0 to their dataset
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
ACCEPT 1representative citing papers
citing papers explorer
-
HPLT 3.0: Very Large-Scale Multilingual Resources for LLMs and MT. Mono- and Bi-lingual Data, Multilingual Evaluation, and Pre-Trained Models
HPLT 3.0 releases the largest open multilingual LLM pre-training dataset at 30 trillion tokens for nearly 200 languages, with quality annotations, evaluation benchmarks for nine languages, and pre-trained models.