AI Paper Atlas

Research map / Architectures and efficiency / Quantization

Qronos: Correcting the Past by Shaping the Future... in Post-Training Quantization

Shihao Zhang, Haoyu Zhang, Ian Colbert, Rayan Saab

ICLR 2026 · Architectures and efficiency › Quantization · approach: Post-training quantization of ViTs · 15 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
replace OPTQ rounding with Qronos (corrects activation/previous-layer mismatch)HIP+MagR 2-bit weight-only · WikiText2, Llama3.2-1Bperplexity: 24.6 -> 17.8Improved
replace OPTQ rounding with QronosHIP+MagR 1.58-bit weight-only · WikiText2, Llama3-8Bperplexity: 43.3 -> 18.0Improved
use Qronos instead of GPTAQ at 2-bit with unfavorable grid beta=1 (ablation)HIP+MagR · WikiText2, Llama3-8Bperplexity: 16.4 -> 12.4Improved
grid scale beta=0.8 instead of beta=1 for 2-bit (ablation)Qronos HIP+MagR · WikiText2, Llama3-8Bperplexity: 12.4 -> 9.3Improved
replace OPTQ with Qronos in W3 HIP weight-onlyLlama3 · WikiText2, 8Bperplexity: 8.6 -> 7.8Improved
replace GPTAQ with Qronos at W4 HIPLlama3-8B · 0-shot avg, W4 HIPaccuracy: 72.4 -> 71.0Hurt
replace OPTQ with Qronos in W4A4KV4 SpinQuantLlama3-3B · WikiText2perplexity: 15.0 -> 11.1Improved
replace GPFQ with Qronos in W3A3 QuaRotLlama3-8B · WikiText2perplexity: 27.9 -> 20.4Improved
add Qronos instead of GPTAQ in W4A4KV4 QuaRot at 8BLlama3-8B · WikiText2perplexity: 9.3 -> 9.3No clear effect
use Qronos instead of OPTQ on instruct Llama3 W3 no transformLlama3-8B-Instruct · WikiText2perplexity: 12.2 -> 10.6Improved
SmoothQuant gamma 0.3 vs 0.8 (RTN, 4-bit) (ablation)Llama3.2-1B-Instruct · WikiText2perplexity: 3e4 -> 18.6Improved
MagR theta 0.01 vs 0.001 (ablation)Llama3.2-1B-Instruct RTN 4-bit · WikiText2perplexity: 105.0 -> 25.4Improved

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