QSLAW (LLM quantization): cited as a baseline — critiqued by newer work, not yet beaten on a benchmark here. 2 paper(s) critique it, 0 beat it on benchmarks — #48 of 80 most-superseded. Sub-problem: cluster led by GPTAQ. Newer alternatives in the same sub-problem include SplitQ, MARR, Quantization-aware Integrated Gradients (QIG), SPEED-Q, Quant-dLLM.

LLM quantization

QSLAW

Superseded baseline#48 of 80 most-superseded

Cited as a baseline — critiqued by newer work, not yet beaten on a benchmark here

2 papers critique it · 0 beat it on benchmarks

What papers say

Verbatim critique sentences, each from a paper that cites QSLAW as a baseline.

these methods either require expensive parameter fine-tuning xie2024qslaw, specialized manipulation at inference time yu2025mquant, or rely on a suboptimal grid search li2025mbq, failing to offer an efficient and effective solution for both calibration and inference

— VLMQ: Efficient Post-Training Quantization for Large Vision-Language Models via Hessian Augmentation

“only quantizes the language component to 4-bit and leaves the vision module at its original precision (i.e., FP16).”

— SPEED-Q: Staged Processing with Enhanced Distillation towards Efficient Low-bit On-device VLM Quantization

What to use instead

Recent methods in the same sub-problem, not yet superseded in the knowledge base — arXiv benchmark leaders, not vetted production recommendations.