Pith. sign in

REVIEW 18 cited by

Prompt Cache: Modular Attention Reuse for Low-Latency Inference

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 2311.04934 v2 pith:XZSBIJGQ submitted 2023-11-07 cs.CL cs.AI

Prompt Cache: Modular Attention Reuse for Low-Latency Inference

classification cs.CL cs.AI
keywords promptcacheinferenceattentionpromptssegmentsreusestates
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

We present Prompt Cache, an approach for accelerating inference for large language models (LLM) by reusing attention states across different LLM prompts. Many input prompts have overlapping text segments, such as system messages, prompt templates, and documents provided for context. Our key insight is that by precomputing and storing the attention states of these frequently occurring text segments on the inference server, we can efficiently reuse them when these segments appear in user prompts. Prompt Cache employs a schema to explicitly define such reusable text segments, called prompt modules. The schema ensures positional accuracy during attention state reuse and provides users with an interface to access cached states in their prompt. Using a prototype implementation, we evaluate Prompt Cache across several LLMs. We show that Prompt Cache significantly reduce latency in time-to-first-token, especially for longer prompts such as document-based question answering and recommendations. The improvements range from 8x for GPU-based inference to 60x for CPU-based inference, all while maintaining output accuracy and without the need for model parameter modifications.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 18 Pith papers

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

  1. Pixels for Programs? A Cross-Provider Case Study of Input-Token Accounting for Source Code as Text and Images

    cs.SE 2026-07 conditional novelty 7.0

    Across 675 paired API calls, image input-token reductions are 86.5% (Anthropic), 80.6% (OpenAI), and 75.8% (Gemini) under token-volume weighting, with Gemini images costing far more than text below 200 lines.

  2. CRAwLeR -- Cross-Reference Aware Legal Retrieval

    cs.IR 2026-06 unverdicted novelty 7.0

    CRAwLeR creates and validates Danish and Polish benchmarks for cross-reference-aware legal chunk retrieval, where current methods reach only 55-59% Recall@10 and the gap is attributed to the contextualizing LLM.

  3. Execution-State Capsules: Graph-Bound Execution-State Checkpoint and Restore for Low-Latency, Small-Batch, On-Device Physical-AI Serving

    cs.LG 2026-06 unverdicted novelty 7.0

    Execution-state capsules enable graph-bound full-state checkpointing and sub-millisecond restore for LLMs including KV and recurrent states, yielding 3.9x-27x TTFT speedups in on-device physical-AI serving.

  4. Move the Query, Not the Cache: Characterizing Cross-Instance Latent Attention Redistribution Across GPU Fabrics

    cs.DC 2026-05 unverdicted novelty 7.0

    On a real multi-node H100 cluster the authors show that for MLA, routing the ~1 KB compressed query row is cheaper than moving cache chunks and supply a topology-aware cost model accurate to ~7% on IBGDA fabrics.

  5. Continuous Semantic Caching for Low-Cost LLM Serving

    cs.LG 2026-04 unverdicted novelty 7.0

    Establishes the first rigorous framework for continuous semantic caching of LLM responses using ε-net discretization and kernel ridge regression, with sublinear regret bounds.

  6. PrefixWall: Mitigating Prefix Caching Side Channels in Shared LLM Systems

    cs.CR 2026-03 unverdicted novelty 7.0

    PrefixWall mitigates APC side channels in multi-tenant LLM systems via selective prefix isolation, delivering up to 70% higher cache reuse and 30% lower latency than full-isolation baselines.

  7. Efficient Remote KV Cache Reuse with GPU-native Video Codec

    cs.DC 2026-02 conditional novelty 7.0

    KVCodec uses GPU-native video codecs and pipelined fetching to compress and transmit KV caches, delivering up to 3.51x faster TTFT than prior methods while preserving accuracy.

  8. Smarter and Cheaper at Once: Byte-Exact KV-Cache Grafting Turns a Frozen Small Model into a Verified-Knowledge Flywheel

    cs.CL 2026-07 conditional novelty 6.0

    Verified solutions can be stored as byte-exact KV-cache blocks and grafted into a frozen LLM, yielding an AIME 2025 gain (80.0% to 93.3%) and a 6,574x cut in decode tokens on recurring problems.

  9. KernelSight-LM: A Kernel-Level LLM Inference Simulator

    cs.PF 2026-06 unverdicted novelty 6.0

    KernelSight-LM simulates LLM inference at kernel granularity with cross-generation (12.1% per-kernel error) and target-measured (3.8% error) tiers, yielding end-to-end median errors of 15.4%/12.8%/3.0% and 14.3%/6.2%/...

  10. KernelSight-LM: A Kernel-Level LLM Inference Simulator

    cs.PF 2026-06 unverdicted novelty 6.0

    KernelSight-LM simulates token-level LLM inference to predict per-kernel latencies and end-to-end metrics (TTFT, TPOT, throughput) with 12.1% and 3.8% kernel errors in cross-generation and target-measured tiers.

  11. A Deterministic Control Plane for LLM Coding Agents

    cs.SE 2026-06 unverdicted novelty 6.0

    Prevalence study of 10k GitHub repos identifies duplication, low revision, and missing permissions in LLM agent configs; proposes Rel(AI)Build control plane with content addressing, tiered permissions, state machine, ...

  12. SGLang: Efficient Execution of Structured Language Model Programs

    cs.AI 2023-12 conditional novelty 6.0

    SGLang is a new system that speeds up structured LLM programs by up to 6.4x using RadixAttention for KV cache reuse and compressed finite state machines for output decoding.

  13. Visko Orbis 1.0: A Live Model for Real-Time Interactive Long Video Generation

    cs.CV 2026-07 conditional novelty 5.0

    A chunk-wise streaming video model with bounded multi-scale memory and streaming 4K upscaling reports real-time interactive long-form generation and top Arena preference/stability scores.

  14. MiniPIC: Flexible Position-Independent Caching in <100LOC

    cs.LG 2026-06 unverdicted novelty 5.0

    MiniPIC enables multiple position-independent caching methods inside vLLM via unrotated KV storage, per-request RoPE application, and three primitives, delivering 49% prefill throughput gains and up to 100x lower cach...

  15. SIFT: Selective-Index For Fast Compute of RAG Prefill by Exploiting Attention Invariance

    cs.AI 2026-06 unverdicted novelty 5.0

    SIFT precomputes selective attention indices via local and cross-attention invariance to speed RAG prefill 1.71x while keeping accuracy within 1% of full recompute, storing only bit vectors 24,000x smaller than KV tensors.

  16. HieraSparse: Hierarchical Semi-Structured Sparse KV Attention

    cs.DC 2026-04 unverdicted novelty 5.0

    HieraSparse delivers a hierarchical semi-structured sparse KV attention system that achieves 1.2x KV compression and 4.57x decode attention speedup versus prior unstructured sparsity methods at equivalent sparsity, pl...

  17. CacheProbe: Auditing Prompt Cache Isolation in Gateway APIs

    cs.CR 2026-05 unverdicted novelty 4.0

    Audits if OpenRouter's gateway architecture with shared credentials inadvertently enables cross-user prompt cache leakage despite provider isolation guarantees.

  18. Rethinking LLMOps for Fraud and AML: Building a Compliance-Grade LLM Serving Stack

    cs.AI 2026-05 unverdicted novelty 4.0

    Workload-aware optimizations for LLM serving in AML and fraud detection yield substantial gains in throughput, latency, and GPU utilization on synthetic compliance prompts.