Channel-wise scaling applied before Hadamard rotation reduces 4-bit layer-wise quantization error in LLaMA2-7B, especially on down-projection layers with massive activation outliers.
Llama 2: Open Foundation and Fine-Tuned Chat Models
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
-
Turning LLM Activations Quantization-Friendly
Channel-wise scaling applied before Hadamard rotation reduces 4-bit layer-wise quantization error in LLaMA2-7B, especially on down-projection layers with massive activation outliers.