Pith. sign in

REVIEW 6 cited by

KV Cache Compression, But What Must We Give in Return? A Comprehensive Benchmark of Long Context Capable Approaches

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2407.01527 v2 pith:ISEHAEA4 submitted 2024-07-01 cs.CL

classification cs.CL
keywords longcontextapproachescachellmsmodelsworkcapability
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
read the original abstract

Long context capability is a crucial competency for large language models (LLMs) as it mitigates the human struggle to digest long-form texts. This capability enables complex task-solving scenarios such as book summarization, code assistance, and many more tasks that are traditionally manpower-intensive. However, transformer-based LLMs face significant challenges with long context input due to the growing size of the KV cache and the intrinsic complexity of attending to extended inputs; where multiple schools of efficiency-driven approaches - such as KV cache quantization, token dropping, prompt compression, linear-time sequence models, and hybrid architectures - have been proposed to produce efficient yet long context-capable models. Despite these advancements, no existing work has comprehensively benchmarked these methods in a reasonably aligned environment. In this work, we fill this gap by providing a taxonomy of current methods and evaluating 10+ state-of-the-art approaches across seven categories of long context tasks. Our work reveals numerous previously unknown phenomena and offers insights - as well as a friendly workbench - for the future development of long context-capable LLMs. The source code is available at https://github.com/henryzhongsc/longctx_bench.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 6 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. CommVQ: Commutative Vector Quantization for KV Cache Compression

    cs.CL 2025-06 conditional novelty 7.0 of 10

    CommVQ compresses LLM key-value caches to roughly one bit per value using additive quantization with a codebook designed to commute with rotary position embeddings, preserving most benchmark accuracy.

  2. Deterministic Inference across Tensor Parallel Sizes That Eliminates Training-Inference Mismatch

    cs.LG 2025-11 conditional novelty 6.0 of 10

    Tree-Based Invariant Kernels fix the floating-point reduction order across GPUs, making LLM logits and sampled tokens bitwise identical for tensor-parallel sizes 1/2/4/8 and exactly matching vLLM (TP=4) with FSDP (TP=1).

  3. SALE : Low-bit Estimation for Efficient Sparse Attention in Long-context LLM Prefilling

    cs.LG 2025-05 conditional novelty 6.0 of 10

    SALE is a training-free sparse attention method that uses 4-bit quantized query-key estimates and a relative attention score to skip unimportant blocks, achieving over 3.36x prefill speedup on 64K+ token contexts with...

  4. ProxSparse: Regularized Learning of Semi-Structured Sparsity Masks for Pretrained LLMs

    cs.LG 2025-02 conditional novelty 6.0 of 10

    ProxSparse learns 2:4 semi-structured sparsity masks for pretrained LLMs via regularized proximal-gradient optimization, outperforming heuristic baselines on seven models.

  5. A Survey on Large Language Model Acceleration based on KV Cache Management

    cs.AI 2024-12 conditional novelty 4.0 of 10

    A survey that classifies KV cache management techniques for faster LLM inference into token-level, model-level, and system-level categories, with benchmark resources.

  6. A Survey of RWKV

    cs.CL 2024-12 conditional novelty 3.0 of 10

    A review of the RWKV architecture, its versions, applications, benchmarks, and open-source ecosystem; it presents no new experimental results.

Pith tools