Study Shows LLM Attention Values Better Capture Sentence Meaning Than Hidden States
Researchers have proposed a new method called Value Aggregation (VA) that uses attention value vectors from large language models to generate sentence embeddings, outperforming existing training-free approaches. Most current LLM-based embedding methods rely on final-layer hidden states, which are optimized for next-token prediction rather than capturing global sentence semantics. The findings suggest a more computationally efficient path to high-quality sentence representations without requiring additional model training.
A paper posted to arXiv introduces Value Aggregation (VA), a training-free method for deriving sentence embeddings from large language models by pooling attention value vectors across multiple layers and token indices, rather than relying on the commonly used final-layer hidden states. The authors argue that hidden states are optimized for next-token prediction and therefore poorly suited for capturing sentence-level semantic meaning. In benchmarks, VA outperforms other training-free LLM-based embedding methods and matches or surpasses MetaEOL, an ensemble-based approach that carries higher computational cost. The paper further introduces a refined variant, Aligned Weighted VA (AlignedWVA), which interprets layer attention outputs as aligned weighted value vectors using the last token's attention scores as weights and the output projection matrix to align them with the LLM residual stream; this method achieves state-of-the-art performance among training-free embeddings, outperforming MetaEOL by a substantial margin. The authors also demonstrate that fine-tuning the Value Aggregation approach holds promise for building strong embedding models with supervised training.
What's missing
The paper is a preprint and has not yet undergone formal peer review.
What different sources said
- arXiv cs.CLCenter
LLM-based Embeddings: Attention Values Encode Sentence Semantics Better Than Hidden States
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.