REVIEW 17 cited by
D2O: Dynamic Discriminative Operations for Efficient Long-Context Inference of Large Language Models
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
D2O: Dynamic Discriminative Operations for Efficient Long-Context Inference of Large Language Models
read the original abstract
Generative inference in Large Language Models (LLMs) is impeded by the growing memory demands of Key-Value (KV) cache, especially for longer sequences. Traditional KV cache eviction strategies, which discard less critical KV pairs based on attention scores, often degrade generation quality, leading to issues such as context loss or hallucinations. In this work, we introduce Dynamic Discriminative Operations (D2O), a KV cache compression method that optimizes KV cache size dynamically and discriminatively at two levels without fine-tuning, while preserving essential context. At layer level, D2O leverages the varying densities of attention weights between shallow and deep layers to dynamically determine which layers should avoid excessive eviction via a novel dynamic allocation strategy to minimize information loss. At token level, D2O incorporates a compensation mechanism that maintains a similarity threshold to re-discriminate the importance of currently discarded tokens, determining whether they should be recalled and merged with similar tokens. We conduct experiments on various benchmarks and LLM architectures. Our results show that D2O not only achieves significant memory savings and enhances inference throughput by more than 3$\times$ but also maintains high-quality long-text generation.
Forward citations
Cited by 17 Pith papers
-
HACK++: Towards More Effective Head-Aware Key-Value Compression for Efficient Visual Autoregressive Modeling
HACK++ is a head-aware KV cache compression framework for VAR models that decouples current-scale attention from historical cache under adaptive per-head budgets to achieve near-lossless generation at 30% attention an...
-
LongLive-RAG: A General Retrieval-Augmented Framework for Long Video Generation
LongLive-RAG formulates long video generation as retrieval-augmented generation by treating self-generated latents as a dynamic searchable history and adding a Window Temporal Delta Loss for better retrieval.
-
Head-Aware Key-Value Compression for Efficient Autoregressive Image Generation
HeadKV compresses KV cache for autoregressive image generation via head-aware budget allocation, early head-type identification from consistent patterns, and stratified token eviction.
-
STAC: Plug-and-Play Spatio-Temporal Aware Cache Compression for Streaming 3D Reconstruction
STAC compresses KV caches in streaming 3D reconstruction transformers via temporal token preservation with decayed attention, spatial voxel compression, and chunked multi-frame optimization, delivering 10x memory redu...
-
PhyCheck: Fine-Grained Evidence-Grounded Dataset for Physical Law Understanding in Video-LLMs
PhyCheck is a 69,825-pair video QA benchmark that tests and improves Video-LLMs' ability to judge whether events obey physical laws, with fine-grained evidence questions and a context-sensitivity pilot.
-
Language Models Need Sleep: Learning to Self-Modify and Consolidate Memories
Sleep-time Knowledge Seeding plus Dreaming lets LLMs expand capacity, distill fragile in-context memories into stable parameters, and self-improve without human labels.
-
Language Models Need Sleep: Learning to Self-Modify and Consolidate Memories
Language models can use a two-stage sleep process of upward distillation for memory consolidation and RL-based dreaming for unsupervised self-improvement to enable continual learning.
-
Meta-Soft: Leveraging Composable Meta-Tokens for Context-Preserving KV Cache Compression
Meta-Soft dynamically synthesizes targeted soft tokens from a learnable orthogonal meta-library via Gumbel-Softmax selection and uses attention-flow integration to preserve semantic information during KV cache eviction.
-
PRISM: Pareto-Efficient Retrieval over Intent-Aware Structured Memory for Long-Horizon Agents
PRISM achieves higher accuracy than baselines on long-horizon agent tasks at an order-of-magnitude smaller context budget by combining hierarchical bundle search, query-sensitive costing, evidence compression, and ada...
-
PRISM: Pareto-Efficient Retrieval over Intent-Aware Structured Memory for Long-Horizon Agents
PRISM is a new inference-time retrieval system that achieves higher accuracy than baselines on long-horizon agent tasks while using an order of magnitude less context by combining hierarchical graph search, intent-bas...
-
ReST-KV: Robust KV Cache Eviction with Layer-wise Output Reconstruction and Spatial-Temporal Smoothing
ReST-KV formulates KV eviction as layer-wise output reconstruction optimization with spatial-temporal smoothing, outperforming baselines by 2.58% on LongBench and 15.2% on RULER while cutting decoding latency by 10.61...
-
Reformulating KV Cache Eviction Problem for Long-Context LLM Inference
LaProx reformulates KV cache eviction as an output-aware matrix approximation, enabling a unified global token selection strategy that preserves LLM performance at 5% cache size across long-context benchmarks.
-
Unifying Sparse Attention with Hierarchical Memory for Scalable Long-Context LLM Serving
SPIN co-designs sparse attention with hierarchical memory to achieve 1.66-5.66x higher throughput, 7-9x lower TTFT, and up to 58% lower TPOT than vLLM and original sparse implementations.
-
RetroInfer: A Vector Storage Engine for Scalable Long-Context LLM Inference
RetroInfer introduces the wave index and wave buffer to realize sparse KV-cache attention for long-context LLM inference with up to 4.4X throughput gains while matching full-attention accuracy.
-
Information-Aware KV Cache Compression for Long Reasoning
InfoKV is an entropy-aware KV cache compression framework that integrates token-level predictive uncertainty and layer-wise representation evolution with attention scores to improve performance on long-context reasoni...
-
A Simple Plug-in for Improving Eviction-Based KV Cache Compression
VECTOR augments eviction-based KV cache compression with three-way token routing that combines importance scoring and offline regression-based reconstructability estimation to improve quality at high compression ratios.
-
Meta-Soft: Leveraging Composable Meta-Tokens for Context-Preserving KV Cache Compression
Meta-Soft dynamically synthesizes targeted soft tokens from a learnable meta-library using Gumbel-Softmax and applies attention-flow integration to compress KV cache while attempting to preserve evicted context information.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.