AI Paper Atlas

Research map / Architectures and efficiency / Quantization

STLA: Spatiotemporal Lookahead Alignment for Post-Training Quantization

Zuqi Zhang, Chenghe Sun, Xiangyi Chu, Wei-Han Yu, Ka-Fai Un, Rui P. Martins, Pui-In Mak, Jiawei Xu

ICML 2026 · Architectures and efficiency › Quantization · approach: Post-training quantization of LLMs

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
couple learning-based rounding with compensation (channel-wise) (ablation)STLA · OPT-125M C4, channel-wiseperplexity: 31.21 -> 30.69Improved
couple learning and compensation (channel-wise, local loss) (ablation)STLA · OPT-125M C4, channel-wiseperplexity: 31.21 -> 32.63Hurt
use lookahead global objective instead of local loss (ablation)STLA coupled · OPT-125M C4, channel-wiseperplexity: 32.63 -> 30.69Improved
use lookahead global objective instead of local loss (ablation)STLA coupled, group 256 · OPT-125M WikiText-2perplexity: 35.86 -> 32.00Improved
use global objective instead of local loss (ablation)STLA coupled, group 256 · LLaMA2-7B WikiText-2perplexity: 6.11 -> 6.44Hurt
add SpinCluster ordering (ablation)STLA coupled, global loss, group 256 · OPT-125M C4perplexity: 29.86 -> 29.26Improved
add SpinCluster ordering (ablation)STLA coupled, global loss, group 256 · OPT-125M WikiText-2perplexity: 32.00 -> 31.01Improved
add SpinCluster ordering (ablation)STLA coupled, global loss, group 256 · LLaMA2-7B WikiText-2perplexity: 6.44 -> 6.09Improved
couple learning and compensation in cluster-wise fashion (ablation)STLA · LLaMA2-7B GPU runtimequantization time: 2.3 HR -> 32.8 MINImproved
use group-wise (size 256) instead of channel-wise quantization (ablation)STLA · OPT-125M C4perplexity: 30.69 -> 29.86Improved
replace GPTAQ with STLA in the rotation pipelineQuaRot · LLaMA2-7B W2A4 WikiText-2perplexity: 16.90 -> 11.42Improved
replace GPTAQ with STLA in the rotation pipelineSpinQuant · LLaMA3-8B W3A4 WikiText-2perplexity: 9.30 -> 8.91Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy