arXiv AI
Aug 13
LinearKV: One Cached State Suffices for Position-Independent Caching in Hybrid LLMs
arXiv:2608. 11231v1 Announce Type: new Abstract: LLM serving is increasingly accelerated by position-independent caching (PIC).
Regularize or Localize: When Training-Time KV-Cache Geometry Pays Under Quantization
arXiv:2607. 17019v1 Announce Type: new Abstract: We study whether \sigreg -- LeJEPA's anti-collapse objective -- can reshape representations during standard autoregressive language-model pretraining, and when the resulting geometry helps \kv-cache quantization.
Related stories
arXiv Computation and Language
Sep 7
Quality Recovery for Quantized KV Caches via Low-Rank Attention Adaptation
The paper investigates how to recover language model quality lost when using low‑bit key–value caches for autoregressive decoding. By keeping the quantizer fixed and distilling the full‑precision cache behavior into low‑rank Q/K/V projection updates, the authors demonstrate that 4‑bit affine‑cache adapters recover roughly 54 % of the perplexity gap on TinyLlama‑1.1B and 76 % on Gemma‑4‑12B, while preserving most long‑context retrieval. Even a 2‑bit rank–token sweep can dramatically reduce TinyLlama’s perplexity, though it only partially restores retrieval performance.
arXiv Machine Learning
Jul 7
Variable Bit-width Quantization: Learning Per-Group Precision for "Bigger-but-Smaller" Language Models
arXiv:2607. 02893v1 Announce Type: new Abstract: Low-bit quantization shrinks language models but treats precision as a single global hyper-parameter: every weight uses the same bit-width.
arXiv Machine Learning
Sep 4
VestigeKV: The NoPE-MLA KV Cache Carries Its Own Eviction Signal in a Vestigial Branch
VestigeKV is a new KV‑cache technique that uses a 64‑dimensional vestigial branch—originally a RoPE component repurposed during NoPE training—as a query‑independent eviction signal. By reading only 11 % of each cache row, the method partitions the cache into an attended tier (top‑m rows) and an archive tier (all other rows), which is GPU‑resident and never deleted. The approach achieves near‑perfect retrieval (1.00 at 8×, 0.92 at 32×) without any training, quantization, or changes to weights or kernels.
arXiv Machine Learning
Aug 4
RestoreKV: Recovering Full-Cache Behavior Under Aggressive Query-Agnostic KV Cache Eviction
arXiv:2608. 01247v1 Announce Type: cross Abstract: Query-agnostic KV cache eviction compresses a context once and reuses the resulting cache for arbitrary future queries, but performance can collapse under tight budgets.
arXiv Computation and Language
Sep 10
VestigeKV: The NoPE-MLA KV Cache Carries Its Own Sparse-Attention Signal in a Vestigial Branch
arXiv:2609.03949v2 Announce Type: replace-cross Abstract: A long-lived KV cache must be compressed before the queries that will read it exist. Selection by observed attention collapses there: on a No...
Hugging Face Trending Papers
Aug 2
RestoreKV: Recovering Full-Cache Behavior Under Aggressive Query-Agnostic KV Cache Eviction
Query-agnostic KV cache eviction compresses a context once and reuses the resulting cache for arbitrary future queries, but performance can collapse under tight budgets. Existing methods primarily improve which original KV pairs are retained.
arXiv AI
Aug 26
Elastic KV Cache for LLM Serving:A Working Reclamation Mechanism, and Why Chunked Prefill Already Closes the Gap
The paper introduces an elastic key‑value (KV) cache for large language model (LLM) serving that dynamically reclaims a pre‑allocated reserve during decode‑heavy phases and restores it before prefill, using a userspace CUDA virtual‑memory trick that requires no driver changes. The authors implement this mechanism, test it under realistic workloads, and find that it offers only marginal benefits—about a 1 % difference in time‑to‑first‑token for large prefill chunks—and that simpler strategies such as lowering the maximum batch size can achieve similar results. The study also notes that the reserve’s impact diminishes with higher tensor‑parallelism levels. whyItMatters":"The work demonstrates that a dynamic KV cache reclamation strategy can be implemented without driver patches and that its practical benefits are limited, guiding future LLM serving optimizations toward simpler approaches."
arXiv Machine Learning
Jul 2
GSRQ: Gain-Shape Residual Quantization for Sub-1-bit KV Cache
arXiv:2607. 01065v1 Announce Type: new Abstract: The deployment of Large Language Models (LLMs) with extended context windows is increasingly constrained by the linear growth of Key-Value (KV) cache memory.
arXiv Machine Learning
Jun 8
Data-Constrained Language Model Pretraining: Improved Regularization and Scaling Laws
arXiv:2606. 06888v1 Announce Type: new Abstract: Classical scaling laws for language model pretraining balance model size against training dataset size under a fixed compute budget, assuming abundant data and a single pass over the corpus.
arXiv AI
Sep 16
Shared-Prefix KV Reuse Across Standard LoRA Adapters: Quality and Serving Tradeoffs
arXiv:2609.17109v1 Announce Type: new Abstract: A common small-model deployment runs one shared backbone with several LoRA specialists that answer over the same context. Serving them naively re-prefi...
arXiv Machine Learning
Jun 9
SpectrumKV: Per-Token Mixed-Precision KV Cache Transfer for Prefill-Decode Disaggregated LLM Serving
arXiv:2606. 08635v1 Announce Type: new Abstract: Prefill-decode (PD) disaggregation decouples prompt processing from token generation, but it also turns the key-value (KV) cache into a network payload.