Pith. sign in

REVIEW 6 cited by

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 2407.14057 v1 pith:DVJ4QDEC submitted 2024-07-19 cs.CL cs.AIcs.LG

LazyLLM: Dynamic Token Pruning for Efficient Long Context LLM Inference

classification cs.CL cs.AIcs.LG
keywords lazyllmprefillingstagetokentokensfirstgenerategeneration
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

The inference of transformer-based large language models consists of two sequential stages: 1) a prefilling stage to compute the KV cache of prompts and generate the first token, and 2) a decoding stage to generate subsequent tokens. For long prompts, the KV cache must be computed for all tokens during the prefilling stage, which can significantly increase the time needed to generate the first token. Consequently, the prefilling stage may become a bottleneck in the generation process. An open question remains whether all prompt tokens are essential for generating the first token. To answer this, we introduce a novel method, LazyLLM, that selectively computes the KV for tokens important for the next token prediction in both the prefilling and decoding stages. Contrary to static pruning approaches that prune the prompt at once, LazyLLM allows language models to dynamically select different subsets of tokens from the context in different generation steps, even though they might be pruned in previous steps. Extensive experiments on standard datasets across various tasks demonstrate that LazyLLM is a generic method that can be seamlessly integrated with existing language models to significantly accelerate the generation without fine-tuning. For instance, in the multi-document question-answering task, LazyLLM accelerates the prefilling stage of the LLama 2 7B model by 2.34x while maintaining accuracy.

discussion (0)

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

Forward citations

Cited by 6 Pith papers

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

  1. Long Context Pre-Training with Lighthouse Attention

    cs.CL 2026-05 conditional novelty 7.0

    Lighthouse Attention enables faster long-context pre-training via gradient-free symmetrical hierarchical compression of QKV while preserving causality, followed by a short full-attention recovery that yields lower los...

  2. 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.

  3. ProactiveLLM: Learning Active Interaction for Streaming Large Language Models

    cs.CL 2026-05 unverdicted novelty 6.0

    ProactiveLLM enables active interaction in streaming LLMs by learning semantic sufficiency cues from partial inputs through mask-based modeling and synchronized privileged self-distillation without external supervision.

  4. UniPrefill: Universal Long-Context Prefill Acceleration via Block-wise Dynamic Sparsification

    cs.CL 2026-05 unverdicted novelty 6.0

    UniPrefill accelerates LLM prefill via block-wise dynamic sparsification, achieving up to 2.1x TTFT speedup while supporting hybrid architectures and native vLLM continuous batching.

  5. HISA: Efficient Hierarchical Indexing for Fine-Grained Sparse Attention

    cs.LG 2026-03 unverdicted novelty 6.0

    HISA speeds up fine-grained sparse attention indexers via block-then-token hierarchy, delivering substantial speedups at 64K context with no training and quality matching the original DSA on long-context benchmarks.

  6. Towards Efficient Large Language Model Serving: A Survey on System-Aware KV Cache Optimization

    cs.LG 2026-07 accept novelty 4.0

    A survey organizing serving-time KV cache optimization techniques into temporal, spatial, and structural system behaviors, analyzing cross-behavior co-design patterns and open challenges.