Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

TORQ: Two-Level Orthogonal Rotation for MXFP4 Quantization

About

As Large Language Models (LLMs) advance toward practical deployment, the Microscaling FP4 (MXFP4) format has emerged as a cornerstone for next-generation low-bit inference, owing to its ability to balance high dynamic range with hardware efficiency. However, directly applying MXFP4 to LLM activation quantization inevitably leads to significant accuracy degradation. In this paper, we theoretically analyze the error structure of MXFP4 activation quantization, revealing that the root cause of this performance drop lies in two structural imbalances between activation distributions and the MXFP4 block floating-point format: (1) extreme inter-block variance imbalance and (2) intra-block codebook utilization imbalance. To address these challenges, we propose TORQ (Two-level Orthogonal Rotation for MXFP4 Quantization), a training-free Post-Training Quantization (PTQ) framework designed to reshape the geometric properties of the activation space through optimal coordinate transformations. At the macroscopic level, TORQ leverages the Schur-Horn theorem to redistribute activation energy via inter-block orthogonal rotation, preventing high-variance blocks from driving up shared scaling factors and thereby preserving the precision of small-magnitude elements. At the microscopic level, TORQ employs maximum-entropy-guided intra-block rotation to alleviate codebook collapse and maximize the MXFP4 codebook's information capacity. Experiments on mainstream LLMs such as LLaMA3 and Qwen3 show that TORQ significantly improves the accuracy of MXFP4 activation quantization compared to existing methods: on Qwen3-32B, the perplexity on WikiText is reduced to 8.43 (vs. 7.61 for BF16), and the average accuracy increases from 38.40% with direct RTN to 73.63% (vs. 74.82% for BF16), substantially narrowing the gap between 4-bit floating-point quantization and full-precision inference.

Zukang Xu, Xing Hu, Dawei Yang• 2026

Related benchmarks

TaskDatasetResultRank
Language ModelingWikiText-2
Perplexity (PPL)3.49
2320
Commonsense ReasoningWinoGrande
Accuracy68.36
1442
Question AnsweringARC Challenge
Accuracy (ARC)55.81
598
Sentence CompletionHellaSwag
Accuracy78.12
364
Mathematical ReasoningMathQA
Accuracy47.94
354
Code GenerationHumanEval
Accuracy61.15
217
Question AnsweringARC Easy
Accuracy77.1
210
Physical ReasoningPIQA
Accuracy77.93
90
Zero-shot EvaluationEvaluation Tasks Zero-shot Aggregate
Avg. Accuracy76.28
74
Showing 9 of 9 rows

Other info

Follow for update