ColBERT-Zero: To Pre-train Or Not To Pre-train ColBERT models

By Antoine Chaffin, Luca Arnaboldi, A. Chatelain, Florent Krzakala. 2026

Subfields: Information Retrieval & Search, Representation & Self-Supervised Learning

Citations in the corpus

• Closing the Operational Gap in Semantic Caching (2026, unknown): “0 ( Santhanam et al., 2022) 0.515 0.403 0.112 0 0.782 Reason-ModernColBERT (Chaffin, 2025b) 0.520 0.382 0.138 0 0.734 ColBERT-Zero (Chaffin et al., 2026) 0.518 0.375 0.143 0 0.724 GTE-ModernColBERT-v1 (Chaffin, 2025a) 0.517 0.354 0.163 0 0.685”
• DenseOn with the LateOn: Fully Open Dense and Late-Interaction Models for Multilingual, Long-Context, and Code Search (2026, unknown): “Finally, we augment supervised fine-tuning data with organic multilingual, long-context, and code data from MIRACL (Zhang et al., 2023), MLDR (Chen et al., 2024), and LateOn-Code (Chaffin, 2026).”