M-Wanda improves one-shot pruning for multilingual LLMs by combining cross-lingual activation variance, activation probability, and correlation-weighted layerwise sparsity, reducing perplexity and boosting downstream tasks relative to Wanda.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
M-Wanda: Improving One-Shot Pruning for Multilingual LLMs
M-Wanda improves one-shot pruning for multilingual LLMs by combining cross-lingual activation variance, activation probability, and correlation-weighted layerwise sparsity, reducing perplexity and boosting downstream tasks relative to Wanda.