infrastructure
The Value Cache Had More Symmetry Than the Model Used
VFold merges cross-layer value states and composes with quantization or key pruning without changing the model architecture.
Summary
VFold merges cross-layer value states and composes with quantization or key pruning without changing the model architecture.
Long contexts make key-value caches a major decoding-memory cost. VFold looks specifically for cross-layer symmetry in value states and merges redundant capacity without requiring an architectural change. The authors report that the method combines with high-ratio quantization or key-cache pruning, reaching compression ratios neither technique achieves alone while adding little decoding overhead. The abstract does not provide one universal compression number, and quality depends on the tested models and tasks. The practical contribution is a composable route to reclaim underused value-cache memory rather than another replacement attention architecture.
Why it matters
VFold merges cross-layer value states and composes with quantization or key pruning without changing the model architecture.
Limits and context
- The abstract does not provide one universal compression number, and quality depends on the tested models and tasks.
Key claims
VFold merges cross-layer value states and composes with quantization or key pruning without changing the model architecture.
Qualification: The abstract does not provide one universal compression number, and quality depends on the tested models and tasks.
Evidence: source-2026-10-11-015
Sources
- arXiv preprint 2610.12338, original v1arXiv / submitting research authors · primary research
Corrections
No corrections have been recorded for this story.