UltraX: Refining Pre-Training Data at Scale with Adaptive Programmatic Editing
As available training data approaches its physical limit, gains from Scaling Laws have begun to diminish. Consequently, improving Large Language Models (LLMs) now depends less on data expansion and more on higher-quality data utilization. However, in the context of large-scale corpora, existing refinement methodologies face significant limitations in quality, efficiency, and reliability: Rule-based approaches are constrained by fixed heuristics and struggle with instance-level variations; LLM-based approaches improve quality but fail to meet the efficiency and reliability requirements of large
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
- Linked via arxiv authorXinlong Zhao →
UltraX: Refining Pre-Training Data at Scale with Adaptive Programmatic Editing
- Linked via arxiv authorDongsheng Liu →
UltraX: Refining Pre-Training Data at Scale with Adaptive Programmatic Editing
- Linked via arxiv authorHengyu Zhao →
UltraX: Refining Pre-Training Data at Scale with Adaptive Programmatic Editing
- Linked via arxiv authorZixuan Fu →
UltraX: Refining Pre-Training Data at Scale with Adaptive Programmatic Editing
- Linked via arxiv authorDezheng Wang →
UltraX: Refining Pre-Training Data at Scale with Adaptive Programmatic Editing
- Linked via arxiv authorJie Cai →
UltraX: Refining Pre-Training Data at Scale with Adaptive Programmatic Editing
- Linked via arxiv authorJie Zhou →
UltraX: Refining Pre-Training Data at Scale with Adaptive Programmatic Editing
- Linked via arxiv authorQiang Ma →
UltraX: Refining Pre-Training Data at Scale with Adaptive Programmatic Editing
- Linked via arxiv authorXuanhe Zhou →
UltraX: Refining Pre-Training Data at Scale with Adaptive Programmatic Editing
- Linked via arxiv authorXu Han →
UltraX: Refining Pre-Training Data at Scale with Adaptive Programmatic Editing
- Linked via arxiv authorYudong Wang →
UltraX: Refining Pre-Training Data at Scale with Adaptive Programmatic Editing
- Linked via arxiv authorZhiyuan Liu →
UltraX: Refining Pre-Training Data at Scale with Adaptive Programmatic Editing