AI Paper Atlas

Research map / Architectures and efficiency / Quantization

LC-QAT: Data-Efficient 2-Bit QAT for LLMs via Linear-Constrained Vector Quantization

Haoyu Wang, Xingyu Yu, Haiyan Zhao, Fengxiang Wang, Xu Han

ICML 2026 · Architectures and efficiency › Quantization · approach: Post-training quantization of LLMs · 0 citations (OpenAlex)

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
initialize QAT from the LC-PTQ result instead of GPTQ or random init (ablation)LC-QAT · Zero-shot QA avg, Qwen-3-8Baverage accuracy: 37.53 (GPTQ init) -> 67.95 (LC-PTQ init); random init 39.51Improved
train end-to-end with LC-QAT instead of PV-Tuning coordinate descent from the same LC-PTQ init (ablation)LC-PTQ model · Zero-shot QA avg, Qwen-3-1.7Baverage accuracy: 55.35 -> 59.75Improved
add integer weight preprocessing (Xavier-aligned proxy weight init) (ablation)LC-QAT · training losstraining lossImproved
use Differentiable Gradient Estimator instead of STE (ablation)LC-QAT · training lossloss spikesImproved
use LC-QAT instead of PV-Tuning (same data)2-bit VQ-QAT · Zero-shot QA avg, Qwen-3-1.7Baverage accuracy: 58.05 -> 61.43Improved
use LC-QAT with 4B tokens instead of ParetoQ with 30B tokensLLaMA-3-8B 2-bit · Zero-shot QA avgaverage accuracy: 72.07 -> 73.42Improved
LC-QAT vs ParetoQ on perplexityLLaMA-3-8B 2-bit · WikiText-2 testperplexity: 8.00 (ParetoQ) vs 9.38 (LC-QAT)Hurt
LC-QAT 1.7B instruct vs BitNet 2B4Tinstruction-following 2-bit model · OpenBookQA/IFEval/MMLU/GSM8K/MATH/HumanEval avgaverage score: 48.07 -> 50.35 average; MMLU 53.17 -> 46.77Mixed
increase training budget from 1B to 10B tokens (ablation)LC-QAT · zero-shot QAaccuracyImproved
vector dimension d=8 instead of d=4 (ablation)LC-PTQ initialization · zero-shot QAaccuracyNo clear effect

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