Stratified sampling preserving teacher score distribution outperforms hard-negative mining as a robust baseline for knowledge distillation in dense retrieval.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.IR 2years
2026 2representative citing papers
citing papers explorer
-
Beyond Hard Negatives: The Importance of Score Distribution in Knowledge Distillation for Dense Retrieval
Stratified sampling preserving teacher score distribution outperforms hard-negative mining as a robust baseline for knowledge distillation in dense retrieval.
- From Tokens to Concepts: Leveraging SAE for SPLADE