Dissecting Outlier Dynamics in LLM NVFP4 Pretraining

• URL: http://arxiv.org/abs/2602.02047v1
• Date: Mon, 02 Feb 2026 12:50:27 GMT
• Title: Dissecting Outlier Dynamics in LLM NVFP4 Pretraining
• Authors: Peijie Dong, Ruibo Fan, Yuechen Tao, Di Mou, Wenhu Hu, Zhenheng Tang, Yinghao Yu, Jiamang Wang, Wenbo Su, Guodong Yang, Liping Zhang, Xiaowen Chu, Baochun Li, Bo Li,
• Abstract summary: This study conducts a longitudinal analysis of outlier dynamics across architecture during NVFP4 pretraining.<n>We find that, compared with Softmax Attention (SA), Linear Attention (LA) reduces per-tensor heavy tails but still exhibits persistent block-level spikes under block quantization.<n>We then develop CHON, an NVFP4 training recipe integrating with post-QK operation protection.
• Score: 46.10969678564592
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Training large language models using 4-bit arithmetic enhances throughput and memory efficiency. Yet, the limited dynamic range of FP4 increases sensitivity to outliers. While NVFP4 mitigates quantization error via hierarchical microscaling, a persistent loss gap remains compared to BF16. This study conducts a longitudinal analysis of outlier dynamics across architecture during NVFP4 pretraining, focusing on where they localize, why they occur, and how they evolve temporally. We find that, compared with Softmax Attention (SA), Linear Attention (LA) reduces per-tensor heavy tails but still exhibits persistent block-level spikes under block quantization. Our analysis attributes outliers to specific architectural components: Softmax in SA, gating in LA, and SwiGLU in FFN, with "post-QK" operations exhibiting higher sensitivity to quantization. Notably, outliers evolve from transient spikes early in training to a small set of persistent hot channels (i.e., channels with persistently large magnitudes) in later stages. Based on these findings, we introduce Hot-Channel Patch (HCP), an online compensation mechanism that identifies hot channels and reinjects residuals using hardware-efficient kernels. We then develop CHON, an NVFP4 training recipe integrating HCP with post-QK operation protection. On GLA-1.3B model trained for 60B tokens, CHON reduces the loss gap to BF16 from 0.94% to 0.58% while maintaining downstream accuracy.

Related papers

• Four Over Six: More Accurate NVFP4 Quantization with Adaptive Block Scaling [13.357423392911036]
  We introduce Four Over Six (4/6), a modification to the NVFP4 quantization algorithm that evaluates two potential scale factors for each block of values.<n>We find that for some blocks, scaling to smaller FP4 values makes the distribution of representable values more uniform.<n>We also find that 4/6 can be easily incorporated into many different post-training quantization methods and generally improves downstream accuracy.
  arXiv  Detail & Related papers  (2025-12-01T18:59:45Z)
• TetraJet-v2: Accurate NVFP4 Training for Large Language Models with Oscillation Suppression and Outlier Control [24.897675627585798]
  Large Language Models (LLMs) training is prohibitively expensive, driving interest in low-precision fully-quantized training (FQT)<n>We introduce TetraJet-v2, an end-to-end 4-bit FQT method that leverages NVFP4 for activations, weights, and gradients in all linear layers.
  arXiv  Detail & Related papers  (2025-10-31T14:57:16Z)
• Pretraining Large Language Models with NVFP4 [53.235038214986865]
  We introduce a novel approach for stable and accurate training of large language models (LLMs) using the NVFP4 format.<n>Our method integrates two-dimensional quantization scheme for consistent representations across both the forward and backward passes.<n>Our results show that the model trained with our NVFP4-based pretraining technique achieves training loss and downstream task accuracies comparable to an FP8 baseline.
  arXiv  Detail & Related papers  (2025-09-29T17:53:17Z)
• Bridging the Gap Between Promise and Performance for Microscaling FP4 Quantization [77.67818998672516]
  We present the first comprehensive study of MXFP4 and NVFP4 for post-training quantization.<n>We introduce Micro-Rotated-GPTQ (MR-GPTQ), a variant of the classic GPTQ quantization algorithm.<n>We show that MR-GPTQ matches or outperforms state-of-the-art accuracy.
  arXiv  Detail & Related papers  (2025-09-27T09:22:21Z)
• Metis: Training LLMs with FP4 Quantization [28.596611044555306]
  Metis is a framework that partitions anisotropic spectra into narrower sub-distributions for independent quantization.<n>On LLaMA-3 8B trained with 100B tokens, Metis enables robust W4A4G4 training with FP4 quantization of weights, activations, and gradients.
  arXiv  Detail & Related papers  (2025-08-30T08:09:08Z)
• FireQ: Fast INT4-FP8 Kernel and RoPE-aware Quantization for LLM Inference Acceleration [1.6127639408026697]
  FireQ is a co-designed PTQ framework and an INT4-FP8 matrix multiplication kernel.<n>FireQ quantizes linear layer weights and key-values to INT4, and activations and queries to FP8.<n>Three-stage pipelining for the prefill phase reduces time-to-first-token in the prefill phase.
  arXiv  Detail & Related papers  (2025-05-27T07:58:35Z)
• Optimizing Large Language Model Training Using FP4 Quantization [73.55459961002371]
  Quantized training presents a promising solution by enabling low-bit arithmetic operations to reduce costs.<n>This work introduces the first FP4 training framework for large language models (LLMs)
  arXiv  Detail & Related papers  (2025-01-28T18:04:50Z)
• SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Models [61.474101404805545]
  Diffusion models can generate high-quality images, but as they scale, rising memory demands and higher latency pose deployment challenges.<n>We propose SVDQuant, a new 4-bit quantization paradigm to overcome this limitation.<n>We reduce the memory usage for the 12B FLUX.1 models by 3.5$times$, achieving 3.0$times$ speedup over the 4-bit weight-only quantization (W4A16) baseline.
  arXiv  Detail & Related papers  (2024-11-07T18:59:58Z)
• Rotated Runtime Smooth: Training-Free Activation Smoother for accurate INT4 inference [54.2589824716527]
  Large language models incur substantial computation and memory movement costs due to their large scale. Existing approaches separate outliers and normal values into two matrices or migrate outliers from activations to weights, suffering from high latency or accuracy degradation. We propose Rotated Smooth (RRS), a plug-and-play activation smoother for quantization, consisting of Smooth and Rotation operation. The proposed method outperforms the state-of-the-art method in the LLaMA and Qwen families and improves WikiText-2 perplexity from 57.33 to 6.66 for INT4 inference.
  arXiv  Detail & Related papers  (2024-09-30T14:59:22Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.