StagQ: Constraint-Driven Multi-Precision Weight Quantization for LLMs

arXiv:2610.05977v1 Announce Type: new Abstract: Serving a large language model (LLM) across a fleet of deployments requires several weight-precision operating points. Multi-precision formats serve them all from one stream whose prefixes are valid lower-precision codes, instead of storing multiple copies. We present StagQ, a multi-precision weight format whose main stream is a 2-bit group-wise affine base followed by a configurable number of 1-bit refinement planes on a dyadic step schedule. Every supported precision is a readable prefix, decoded by an affine map derived from metadata shared across all precisions, with no per-weight lookup. A sparse side record, filled both before and after the grid is fitted, holds out the few weights the grid serves worst. We report two configurations of the encoder. At two bits the cheaper one leads the strongest multi-precision baseline on Llama-3.1-8B, Phi-4, and OLMo-2-7B by 3.1 to 7.0 MMLU points, at a slightly lower logical rate. At three bits it leads on Llama-3.1-8B, leads on Phi-4 at a higher rate, and ties on OLMo-2-7B. At four bits it ties on all three, at a higher rate. In a batch-one matrix-vector product on an NVIDIA A100 GPU, timed on synthetic weights, our kernel is faster than the two baseline kernels in most shape-precision cases.

Related

• Multi-Bitwidth Quantization for LLMs Using Additive Codebooks
• WINDQuant: Weight-Informed Neural Decision-Making for Global Mixed-Precision LLM Quantization
• CoQuant: Joint Weight-Activation Subspace Projection for Mixed-Precision LLMs
• Squeeze10-LLM: Squeezing LLMs' Weights by 10 Times via a Staged Mixed-Precision Quantization Method

Source: arXiv cs.LG | 2026-10-06