Still: New Method for Efficient KV Cache Compression in Large Language Models
Researchers introduced Still, a lightweight neural network module that compresses the key-value cache in large language models during a single forward pass, addressing a major memory bottleneck in long-context inference. The method uses a small per-layer Perceiver trained once against a frozen base model and achieves compression ratios from 8× to 200× while maintaining context quality. This approach is significant because it enables practical deployment of language models on longer contexts while reducing memory requirements, a critical constraint for real-world applications.
Still is a new KV cache compression technique designed to solve the memory bottleneck that limits long-context language model deployment. The method trains a small Perceiver module per layer once against a frozen base model, allowing it to produce compact keys and values in a single forward pass during inference. Testing on Qwen and Gemma models shows Still achieves favorable speed-quality tradeoffs across compression ratios from 8× to 200× and context lengths from 8k to 128k tokens. On the RULER benchmark for long-context evaluation, Still outperforms the strongest baseline by 8-22 points. The approach also supports free-form summarization and can be applied iteratively for extremely long contexts, capabilities unavailable to existing per-context optimization methods. By amortizing the compression cost across inference, Still makes extreme compression ratios practical for real-world deployment.
What's missing
The paper does not discuss potential limitations of the approach, such as how performance degrades with out-of-distribution contexts, computational overhead of the Perceiver module itself, or comparison of wall-clock inference time versus memory savings tradeoffs in practical deployment scenarios.
What different sources said
- arXiv cs.CLCenter
Do Language Models Need Sleep? Offline Recurrence for Improved Online Inference
- VentureBeatCenter
Context compression finally works in production: new research cuts LLM input 16x without the accuracy hit
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.