QLESS quantizes the gradient features used in LESS data selection down to 1 bit without losing much fine-tuning quality, cutting gradient storage up to 16x across several LLM families.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
QLESS: A Quantized Approach for Data Valuation and Selection in Large Language Model Fine-Tuning
QLESS quantizes the gradient features used in LESS data selection down to 1 bit without losing much fine-tuning quality, cutting gradient storage up to 16x across several LLM families.