mlsanityData to 14 Oct 2026

HARP: Hadamard-Preconditioned Adaptive Rotation Processor for Extreme LLM Quantization

arXiv 2605.29843 ↗First posted 28 May 2026cs.LG cs.AIcited by 5

Not on the map. Why?

Abstract (arXiv)

Post-training quantization (PTQ) is essential for deploying LLMs under memory and bandwidth constraints. However, extreme low-bit quantization remains highly sensitive to activation outliers and anisotropic weight curvature. Existing incoherence-based PTQ methods mitigate this issue with fixed randomized Hadamard transforms (RHTs), which improve quantization robustness but cannot adapt the rotated basis to the layer, calibration distribution, or quantizer. We introduce HARP (Hadamard-preconditioned Adaptive Rotation Processor), a learnable structured two-sided orthogonal processor that replaces fixed Hadamard mixing while preserving exact full-precision equivalence. HARP represents each rotation as a product of sparse butterfly-like block-orthogonal stages, supports non-power-of-two dimensions through Mixed-Radix schedules, and initializes to the RHT processor up to a fixed permutation. Fitted only on calibration data, HARP adapts the quantization basis to each layer and backend. Across 2--4-bit settings on Llama models from 1B to 70B, HARP consistently improves perplexity and yields its clearest zero-shot gains at 2 bits; a 2-bit Qwen3-8B experiment shows the same transfer beyond the Llama family. HARP also preserves deployment efficiency: on Llama 2 7B at 2 bits, it reaches 128 tok/s, retaining 90% of RHT throughput (142 tok/s) and running approximately 2.1× faster than FP16 (61 tok/s).

Key facts. Cited by 5 papers in the map (data through 14 Oct 2026). Builds on 26 of its 44 references that are in the map. First posted on arXiv 28 May 2026.

Builds on

26 of 44 references.

References · 26

  1. 2022GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformerscited by 1,919
  2. 2023AWQ: Activation-aware Weight Quantization for LLM Compression and Accelerationcited by 1,298
  3. 2023SparseGPT: Massive Language Models Can Be Accurately Pruned in One-Shotcited by 1,121
  4. 2022LLM.int8(): 8-bit Matrix Multiplication for Transformers at Scalecited by 906
  5. 2024QuaRot: Outlier-Free 4-Bit Inference in Rotated LLMscited by 634
  6. 2022ZeroQuant: Efficient and Affordable Post-Training Quantization for Large-Scale Transformerscited by 526
  7. 2024SpinQuant: LLM quantization with learned rotationscited by 471
  8. 2023QuIP: 2-Bit Quantization of Large Language Models With Guaranteescited by 414
  9. 2024QuIP#: Even Better LLM Quantization with Hadamard Incoherence and Lattice Codebookscited by 378
  10. 2023SpQR: A Sparse-Quantized Representation for Near-Lossless LLM Weight Compressioncited by 367
  11. 2022Optimal Brain Compression: A Framework for Accurate Post-Training Quantization and Pruningcited by 322
  12. 2023SqueezeLLM: Dense-and-Sparse Quantizationcited by 294
14 more
  1. 2024Extreme Compression of Large Language Models via Additive Quantizationcited by 232
  2. 2023Controlling Text-to-Image Diffusion by Orthogonal Finetuningcited by 208
  3. 2023Parameter-Efficient Orthogonal Finetuning via Butterfly Factorizationcited by 114
  4. 2024QTIP: Quantization with Trellises and Incoherence Processingcited by 91
  5. 2024GPTVQ: The Blessing of Dimensionality for LLM Quantizationcited by 82
  6. 2019Learning Fast Algorithms for Linear Transforms Using Butterfly Factorizationscited by 76
  7. 2024PV-Tuning: Beyond Straight-Through Estimation for Extreme LLM Compressioncited by 61
  8. 2025Qronos: Correcting the Past by Shaping the Future... in Post-Training Quantizationcited by 28
  9. 2025Model-Preserving Adaptive Roundingcited by 23
  10. 2025WUSH: Near-Optimal Adaptive Transforms for LLM Quantizationcited by 15
  11. 2025ButterflyQuant: Ultra-low-bit LLM Quantization through Learnable Orthogonal Butterfly Transformscited by 10
  12. 2024The Iterative Optimal Brain Surgeon: Faster Sparse Recovery by Leveraging Second-Order Informationcited by 8
  13. 2025Q-Palette: Fractional-Bit Quantizers Toward Optimal Bit Allocation for Efficient LLM Deploymentcited by 8
  14. 2026WaterSIC: Information-Theoretically (Near) Optimal Linear Layer Quantizationcited by 7

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 5.

5 papers.

  1. 2026Rethinking the Role of Tensor Decompositions in Post-Training LLM Compressioncited by 4
  2. 2026Extreme Low-Bit Inference in Reasoning Models: Failure Modes and Targeted Recoverycited by 1
  3. 2026Transforms for LLM Quantization: The Great Inversion and Format Co-Designcited by 0
  4. 2026RSLM: Training-Free Vector Quantization for Approximate Nearest Neighbor Searchcited by 0
  5. 2026KBBQ: A Predictive Noise Law and the Limits of Spectrum Flattening in FP4 Quantizationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025DartQuant: Efficient Rotational Distribution Calibration for LLM Quantization
  2. 2024DFRot: Achieving Outlier-Free and Massive Activation-Free for Rotated LLMs with Refined Rotation
  3. 2026Qift: Shift-Friendly No-Zero W2 Post-Training Quantization for Rotated W2A4/KV4 LLM Inference
  4. 2026Influence-Inspired Spectral Rotations for Extreme Low-Bit LLM Quantization
  5. 2026AdaHOP: Fast and Accurate Low-Precision Training via Outlier-Pattern-Aware Rotation
  6. 2025Grouped Sequency-arranged Rotation: Optimizing Rotation Transformation for Quantization for Free
  7. 2026HeRo-Q: A General Framework for Stable Low Bit Quantization via Hessian Conditioning
  8. 2024Taming Sensitive Weights : Noise Perturbation Fine-tuning for Robust LLM Quantization