KBBQ: A Predictive Noise Law and the Limits of Spectrum Flattening in FP4 Quantization
Not on the map. Why?
Abstract (arXiv)
We develop a second-order theory of quantization noise in matrix multiplication in which the quantization format is characterized by the variance it assigns to each element. The constant variance profile of integer quantization recovers existing integer-noise theory, while the multiplicative profile of floating-point rounding reduces the data dependence to a scalar, the participation factor kappa, yielding a closed-form signal-to-noise-ratio law. The resulting functional also admits a closed-form upper bound kappa* that no function-preserving linear transform can exceed and that is attained by a recent state-of-the-art method. Building on this analysis, we introduce KBBQ (Kappa-Braked Blockwise Quantization), which parameterizes the extent to which a transform approaches this ceiling. At W4A4, across four base models and two FP4 formats, KBBQ outperforms the prior state of the art without additional deployment-time computation.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 25 of its 34 references that are in the map. First posted on arXiv 8 Sep 2026.
Builds on
25 of 34 references.
References · 25
- 2024The Llama 3 Herd of Modelscited by 14,715
- 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
- 2021Training Verifiers to Solve Math Word Problemscited by 10,109
- 2025Qwen3 Technical Reportcited by 8,820
- 2020Measuring Massive Multitask Language Understandingcited by 7,963
- 2019HellaSwag: Can a Machine Really Finish Your Sentence?cited by 4,100
- 2021Program Synthesis with Large Language Modelscited by 3,637
- 2022GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformerscited by 1,919
- 2022SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Modelscited by 1,386
- 2024QuaRot: Outlier-Free 4-Bit Inference in Rotated LLMscited by 634
- 2024SpinQuant: LLM quantization with learned rotationscited by 471
- 2023QuIP: 2-Bit Quantization of Large Language Models With Guaranteescited by 414
13 more
- 2024QuIP#: Even Better LLM Quantization with Hadamard Incoherence and Lattice Codebookscited by 378
- 2023Microscaling Data Formats for Deep Learningcited by 152
- 2025Pretraining Large Language Models with NVFP4cited by 76
- 2025FP4 All the Way: Fully Quantized Training of LLMscited by 40
- 2025Quartet: Native FP4 Training Can Be Optimal for Large Language Modelscited by 39
- 2025HALO: Hadamard-Assisted Lower-Precision Optimization for LLMscited by 15
- 2025WUSH: Near-Optimal Adaptive Transforms for LLM Quantizationcited by 15
- 2026Full-Stack FP4: Stable LLM Pretraining with Quantized Projections, Optimizers, and Attentioncited by 8
- 2026Dissecting Quantization Error: A Concentration-Alignment Perspectivecited by 5
- 2026HARP: Hadamard-Preconditioned Adaptive Rotation Processor for Extreme LLM Quantizationcited by 5
- 2026TORQ: Two-Level Orthogonal Rotation for MXFP4 Quantizationcited by 3
- 2026KronQ: LLM Quantization via Kronecker-Factored Hessiancited by 3
- 2026ConQuR: Corner Aligned Activation Quantization via Optimized Rotations for LLMscited by 2
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026Optimal Scalar Quantization for Matrix Multiplication: Closed-Form Density and Phase Transition
- 2026A Nuclear-Norm Lower Bound for Dithered Scalar Quantization of Matrix Products
- 2026High-Rate Quantized Matrix Multiplication II
- 2026High-Rate Quantized Matrix Multiplication I
- 2026Influence-Inspired Spectral Rotations for Extreme Low-Bit LLM Quantization
- 2026Transforms for LLM Quantization: The Great Inversion and Format Co-Design
- 2026BCJR-QAT: A Differentiable Relaxation of Trellis-Coded Weight Quantization
- 2026A Note on Scaling in Randomly Rotated Quantization and Its Connection to the CDEF +1 Pythagorean Relation