Weight gradient FP4 quantization drives LLM pretraining divergence, which deterministic Hadamard rotations can stabilize on native MXFP4 hardware.
arXiv preprint arXiv:2603.08747 , year=
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Pretraining large language models with MXFP4 on Native FP4 Hardware
Weight gradient FP4 quantization drives LLM pretraining divergence, which deterministic Hadamard rotations can stabilize on native MXFP4 hardware.