A data composition method that aligns SFT data proportions with a model's detected domain knowledge distribution and dynamically reweights domains by learnable potential improves multi-domain performance versus uniform mixing.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
VersaTune: An Efficient Data Composition Framework for Training Multi-Capability LLMs
A data composition method that aligns SFT data proportions with a model's detected domain knowledge distribution and dynamically reweights domains by learnable potential improves multi-domain performance versus uniform mixing.