CCI4.0: A Bilingual Pretraining Dataset for Enhancing Reasoning in Large Language Models
This addresses the need for high-quality, diverse training data to improve LLM reasoning, though it is incremental as it builds on existing data curation and CoT methods.
The paper tackles the problem of enhancing reasoning in large language models by introducing CCI4.0, a 35 TB bilingual pre-training dataset with curated data and 4.5 billion chain-of-thought templates, resulting in consistent improvements in downstream tasks like math and code reflection.
We introduce CCI4.0, a large-scale bilingual pre-training dataset engineered for superior data quality and diverse human-like reasoning trajectory. CCI4.0 occupies roughly $35$ TB of disk space and comprises two sub-datasets: CCI4.0-M2-Base and CCI4.0-M2-CoT. CCI4.0-M2-Base combines a $5.2$ TB carefully curated Chinese web corpus, a $22.5$ TB English subset from Nemotron-CC, and diverse sources from math, wiki, arxiv, and code. Although these data are mostly sourced from well-processed datasets, the quality standards of various domains are dynamic and require extensive expert experience and labor to process. So, we propose a novel pipeline justifying data quality mainly based on models through two-stage deduplication, multiclassifier quality scoring, and domain-aware fluency filtering. We extract $4.5$ billion pieces of CoT(Chain-of-Thought) templates, named CCI4.0-M2-CoT. Differing from the distillation of CoT from larger models, our proposed staged CoT extraction exemplifies diverse reasoning patterns and significantly decreases the possibility of hallucination. Empirical evaluations demonstrate that LLMs pre-trained in CCI4.0 benefit from cleaner, more reliable training signals, yielding consistent improvements in downstream tasks, especially in math and code reflection tasks. Our results underscore the critical role of rigorous data curation and human thinking templates in advancing LLM performance, shedding some light on automatically processing pretraining corpora.