DarwinLM uses evolutionary search with training-aware offspring selection to prune LLMs, beating ShearedLlama with 5x less post-training data.
Evolving knowledge distillation with large lan- guage models and active learning
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
-
DarwinLM: Evolutionary Structured Pruning of Large Language Models
DarwinLM uses evolutionary search with training-aware offspring selection to prune LLMs, beating ShearedLlama with 5x less post-training data.