Researchers Propose Locality-Aware Redundancy Pruning to Improve LLM Inference Efficiency
Researchers have proposed 'Ghosted Layers,' a training-free module that recovers accuracy lost when entire Transformer decoder blocks are removed from large language models. Layer pruning is a common technique to reduce model size and computational cost, but it creates a mismatch between hidden states and the distributions surviving layers were trained to handle. The method offers a practical path to maintaining efficiency gains from pruning without the cost of retraining.
A preprint posted to arXiv introduces Ghosted Layers, a recovery technique designed to address performance degradation caused by layer pruning in large language models (LLMs). When entire Transformer decoder blocks are removed, the hidden state passed to the next surviving layer no longer matches the distribution that layer was trained on, causing significant accuracy and perplexity losses. Ghosted Layers tackles this by solving a 'boundary activation alignment' problem: it derives a closed-form optimal linear operator from a small calibration dataset to reconstruct the activation discrepancy introduced by the removed layers. The authors argue this solution represents the unconstrained optimum of the alignment objective, whereas prior methods are limited to constrained solutions within narrower operator subspaces. Experiments across multiple LLM architectures and pruning strategies show consistent improvements in accuracy and perplexity over existing training-free baselines, while preserving the computational efficiency gains that make pruning attractive. The method requires no gradient-based retraining, making it computationally lightweight to apply. Code has been made publicly available by the authors.
What's missing
The calibration set size and its sensitivity to domain mismatch are not discussed in the abstract, which are important practical considerations. The method's behavior under aggressive pruning ratios (e.g., removing a large fraction of layers) is not characterized, nor is a comparison to lightweight fine-tuning recovery approaches included.
What different sources said
- arXiv cs.AICenter
LEAP: Learnable End-to-End Adaptive Pruning of Large Language Models
Related
Gut Bacteria Enzyme Found to Break Down Heat-Processed Food Compounds, Producing Novel Biogenic Amines
Researchers have discovered that an enzyme in common gut bacteria can degrade N-epsilon-carboxymethyllysine (CML), a compound formed during thermal food processing, producing previously unknown biogenic amines. The enzyme, ornithine decarboxylase SpeC from enterobacteria, acts on CML and related modified lysine derivatives through a low-level 'underground' catalytic activity. This finding suggests a previously unrecognized communication axis between thermally processed dietary compounds and gut microbial physiology, with potential implications for host health.
Full-Length Gene Sequencing Reveals Two Distinct Bacterial Communities in Black-Legged Ticks Expanding Into Canada
Researchers used Oxford Nanopore full-length 16S rRNA gene sequencing to characterize the microbiome of Ixodes scapularis black-legged ticks collected in Nova Scotia, Canada, distinguishing between tick-adapted bacteria and environmentally acquired bacteria. The study comes as I. scapularis — the primary vector of Lyme disease — is rapidly expanding northward into Canada due to climate change. The findings suggest that environmentally derived bacteria in tick microbiomes are not mere contamination, which has implications for how tick microbiome data is collected and interpreted across surveillance studies.
Study Identifies Metabolic Link Between Cell Envelope Stress and Biofilm Formation in Bacteria
Researchers have discovered that the metabolite acetyl-CoA directly inhibits enzymes that degrade the bacterial signaling molecule c-di-GMP, connecting cell envelope biosynthesis stress to biofilm formation in Pseudomonas aeruginosa. The study found that sub-inhibitory concentrations of antibiotics targeting early peptidoglycan biosynthesis — but not other antibiotic classes — elevate c-di-GMP levels by reducing phosphodiesterase activity, with acetyl-CoA competing for the enzyme active site. Because the relevant enzyme domain is broadly conserved across bacterial species, this checkpoint mechanism may be widespread and could have implications for understanding antibiotic-induced biofilm responses.