cd /news/artificial-intelligence/coec-calibrated-orthogonal-equivalen… · home topics artificial-intelligence article
[ARTICLE · art-108331] src=machinebrief.com ↗ pub= topic=artificial-intelligence verified=true sentiment=· neutral

COEC: Calibrated Orthogonal-Equivalence Compensation for Structured Pruning of Large Language Models

Researchers propose COEC (Calibrated Orthogonal-Equivalence Compensation), a training-free framework that applies alternating left and right orthogonal rotations to retained weights in structured pruning of large language models, improving perplexity on every tested model and zero-shot accuracy in most settings. Experiments on Llama-3, Llama-3.1, and Qwen2.5 families show larger gains at higher sparsity, requiring only second-order statistics from a small calibration set.

read1 min views1 publishedAug 24, 2026

arXiv:2608.21142v1 Announce Type: new Abstract: Structured pruning reduces the size and inference cost of large language models (LLMs) by removing weight columns, but the resulting output error can degrade accuracy. Existing training-free compensation methods use an additive bias or a single orthogonal rotation on the output side of the retained weight. These corrections leave its input singular frame unchanged and therefore limit how the retained weight can adapt after column removal. We propose COEC (Calibrated Orthogonal-Equivalence Compensation), a training-free compensation framework that applies alternating left and right orthogonal rotations to the retained weight. The right rotation is optimized on a reduced Stiefel manifold, while singular values are rescaled using generalized cross-validation to select the regularization strength for each layer. COEC further tempers the calibration Gram matrix to reduce the dominance of high-energy activation directions and introduces an alignment penalty that preserves the geometric relation between adjacent attention projections.All components use second-order statistics from a small calibration set and require neither backpropagation through the LLM nor retraining of the model parameters. COEC is independent of the column pruning criterion and can be applied to multiple structured pruning methods. Experiments on the Llama-3, Llama-3.1, and Qwen2.5 model families across multiple structured sparsity levels show that COEC improves perplexity on every model and zero-shot accuracy in most settings over existing compensation methods, with larger gains at higher sparsity. These results show that post-pruning compensation can recover part of the performance lost to column removal.

── more in #artificial-intelligence 4 stories · sorted by recency
── more on @coec 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/coec-calibrated-orth…] indexed:0 read:1min 2026-08-24 ·