AI Paper Atlas

Research map / Architectures and efficiency / Quantization

SkipCat: Rank-Maximized Low-Rank Compression of Large Language Models via Shared Projection and Block Skipping

Yu-Chen Lu, Sheng-Feng Yu, Hui-Hsien Weng, Pei-Shuo Wang, Yu-Fang Hu, Liang Hung-Chun, Hung-Yueh Chiang, Kai-Chiang Wu

AAAI 2026 · Architectures and efficiency › Quantization · approach: KV cache quantization · 2 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
add Cat (shared low-rank projection) to naive SVD low-rank compression (ablation)SVD-LLM style low-rank compression · WikiText-2 (LLaMA2-7B, 20% compression)Wikitext-2 perplexity: 8.82 -> 7.84Improved
add Skip (block skipping) to naive SVD low-rank compression (ablation)SVD-LLM style low-rank compression · WikiText-2 (LLaMA2-7B, 20% compression)Wikitext-2 perplexity: 8.82 -> 6.71Improved
add Cat on top of Skip (ablation)Skip-only low-rank compression · WikiText-2 (LLaMA2-7B, 20% compression)Wikitext-2 perplexity: 6.71 -> 6.29Improved
add 8-bit HQQ quantization to the compressed model (ablation)SkipCat · WikiText-2 (LLaMA2-7B, 20% compression)Wikitext-2 perplexity: 6.29 -> 6.29No clear effect
add Hadamard transform before 8-bit quantization (ablation)SkipCat 8-bit · WikiText-2 (LLaMA2-7B, 20% compression, 8-bit)Wikitext-2 perplexity: 6.36 -> 6.32Improved
add channel scaling before 8-bit quantization (ablation)SkipCat 8-bit · WikiText-2 (LLaMA2-7B, 20% compression, 8-bit)Wikitext-2 perplexity: 6.36 -> 6.30Improved
add both Hadamard and channel scaling before 8-bit quantization (ablation)SkipCat 8-bit · WikiText-2 (LLaMA2-7B, 20% compression, 8-bit)Wikitext-2 perplexity: 6.36 -> 6.29Improved
add column permutation before block skipping (vs FP16 without it) (ablation)Block skipping · WikiText-2 (LLaMA2-7B, 20% compression, FP16)Wikitext-2 perplexity: Overflow -> 6.29Improved
run block skipping without permutation in BF16 instead of FP16 (10% compression) (ablation)Block skipping without column permutation · WikiText-2 (LLaMA2-7B, 10% compression)Wikitext-2 perplexity: 13.34 -> 31628.76Hurt
add LoRA fine-tuning to SkipCat at 20% compressionSkipCat · Zero-shot 7 tasks (LLaMA2-7B, 20% compression)accuracy drop vs dense (%): 2.20 -> 0.39Improved
use SkipCat instead of SVD-LLM after LoRA fine-tuning at 20% compressionSVD-LLM · Zero-shot 7 tasks (LLaMA2-7B, 20% compression, LoRA)average zero-shot accuracy (%): 51.02 -> 54.41Improved
use SkipCat instead of SVD-LLM at 30% compressionSVD-LLM · Zero-shot 7 tasks (LLaMA2-7B, 30% compression)average zero-shot accuracy (%): 41.25 -> 48.46Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 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