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.
hub
Model tells you where to merge: Adaptive kv cache merging for llms on long-context tasks
14 Pith papers cite this work, alongside 1 external citations. Polarity classification is still indexing.
14
Pith papers citing it
1
external citations · Pith
abstract
How to efficiently serve Large Language Models (LLMs) has become a pressing issue because of their huge computational cost in their autoregressive generation process. To mitigate computational costs, LLMs often employ the KV Cache technique to improve the generation speed. While improving the computational efficiency, the storage requirements of the KV cache are substantial, particularly in long-context scenarios, leading to significant memory consumption. Existing KV cache eviction methods often degrade the performance of LLMs in long-context scenarios due to the information loss introduced by eviction. In this paper, we propose a novel KV cache merging approach, called KVMerger, to achieve adaptive KV cache compression for long-context tasks without significant performance degradation under constrained memory budgets. Our approach is inspired by the intriguing observation that key states exhibit high similarity at the token level within a single sequence. To facilitate merging, we develop an effective yet straightforward merging set identification algorithm to identify suitable KV states for merging. Our merging set identification algorithm stimulates the second observation that KV cache sparsity, from similarity perspective, is independent of the dataset and remains persistent at the model level. Subsequently, we propose a Gaussian kernel weighted merging algorithm to selectively merge all states within each merging set. We conduct extensive experiments to demonstrate the effectiveness of KVMerger for long-context tasks under constrained memory budgets, applying it to models including Llama2-7B-chat and Llama2-13B-chat. Using the LongBench and ZeroScroll benchmarks, we compare our method with other KV cache compression techniques, including H2O and CaM, showing that our method achieves superior performance across tasks with both 50% and 35% KV cache budgets.
hub tools
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
Prefill-Time Intervention (PTI) reduces hallucinations in large vision-language models by applying a one-time modality-aware steering correction to the initial KV cache at the prefill stage rather than during autoregressive decoding.
Transformers need depth scaling as the product of ceil(k/s) and log n terms for k-hop pointer chasing under cache size s, with a conjectured lower bound, proved upper bound via windowed pointer doubling, and an adaptive-oblivious error separation.
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 reduction and 4x faster inference at SOTA quality.
Learned pooling of speech KV caches from an intermediate LLM layer compresses speech to text-level length while matching or exceeding the uncompressed baseline on ASR and entity recognition, with 1.49–2× decoding speedup.
DepthWeave-KV achieves 8.3x KV cache memory reduction with near-full-cache task quality by factorizing key-value states across transformer layers using shared bases and token-adaptive residuals.
EntropyInfer adaptively allocates inference compute using per-head attention entropy for rigid/dynamic classification during prefilling and compresses KV cache with generated tokens, achieving up to 2.39x speedup on long contexts.
Sleep-time Knowledge Seeding plus Dreaming lets LLMs expand capacity, distill fragile in-context memories into stable parameters, and self-improve without human labels.
MomentKV maintains count, key mean, value mean, and value-key covariance over evicted tokens to guide selective eviction and provide a first-order approximation of their attention contribution, outperforming baselines on LongBench and RULER.
Frequency-guided inter-layer KV sharing with logit-aware head routing nearly matches full-cache long-context accuracy at about 3.9× lower peak KV memory.
Document LoRA acts as decoding-time parametric memory that recovers 13-21 ROUGE-L points under heavy KV cache compression in QA, performing best when the base model encodes the document and the adapter is used only at generation with QA supervision.
GRKV applies global ridge regression to KV cache merging for span-based retention in long-context LLMs, claiming to be the only method that improves benchmark performance with minimal overhead.
ReasonCache reuses similar KV cache states across reasoning steps in LRMs via collaborative filtering to boost serving throughput by up to 89.2% while preserving accuracy.
STARC remaps sparse KV caches by semantic clustering for PIM hardware, delivering 19-31% lower attention latency and 19-27% lower energy versus token-wise sparsity, with larger gains under tight KV budgets.
citing papers explorer
-
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.
-
Prefill-Time Intervention for Mitigating Hallucination in Large Vision-Language Models
Prefill-Time Intervention (PTI) reduces hallucinations in large vision-language models by applying a one-time modality-aware steering correction to the initial KV cache at the prefill stage rather than during autoregressive decoding.
-
How Much Cache Does Reasoning Need? Depth-Cache Tradeoffs in KV-Compressed Transformers
Transformers need depth scaling as the product of ceil(k/s) and log n terms for k-hop pointer chasing under cache size s, with a conjectured lower bound, proved upper bound via windowed pointer doubling, and an adaptive-oblivious error separation.
-
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 reduction and 4x faster inference at SOTA quality.
-
Compress the Cache, Not the Speech Embedding: KV Compression for Efficient Speech LLMs
Learned pooling of speech KV caches from an intermediate LLM layer compresses speech to text-level length while matching or exceeding the uncompressed baseline on ASR and entity recognition, with 1.49–2× decoding speedup.
-
DepthWeave-KV: Token-Adaptive Cross-Layer Residual Factorization for Long-Context KV Cache Compression
DepthWeave-KV achieves 8.3x KV cache memory reduction with near-full-cache task quality by factorizing key-value states across transformer layers using shared bases and token-adaptive residuals.
-
From Rigid to Dynamic: Entropy-Guided Adaptive Inference for Long-Context LLMs
EntropyInfer adaptively allocates inference compute using per-head attention entropy for rigid/dynamic classification during prefilling and compresses KV cache with generated tokens, achieving up to 2.39x speedup on long contexts.
-
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.
-
MomentKV: Closing the Directional Gap in KV Cache Eviction for Long-Context Inference
MomentKV maintains count, key mean, value mean, and value-key covariance over evicted tokens to guide selective eviction and provide a first-order approximation of their attention contribution, outperforming baselines on LongBench and RULER.
-
FreqDepthKV: Frequency-Guided Depth Sharing for Robust KV Cache Compression in Long-Context LLM Inference
Frequency-guided inter-layer KV sharing with logit-aware head routing nearly matches full-cache long-context accuracy at about 3.9× lower peak KV memory.
-
Rethinking LoRA Memory Through the Lens of KV Cache Compression
Document LoRA acts as decoding-time parametric memory that recovers 13-21 ROUGE-L points under heavy KV cache compression in QA, performing best when the base model encodes the document and the adapter is used only at generation with QA supervision.
-
GRKV: Global Regression for Training-Free KV Cache Compression in Long-Context LLMs
GRKV applies global ridge regression to KV cache merging for span-based retention in long-context LLMs, claiming to be the only method that improves benchmark performance with minimal overhead.
-
ReasonCache: Accelerating Large Reasoning Model Serving through KV Cache Sharing
ReasonCache reuses similar KV cache states across reasoning steps in LRMs via collaborative filtering to boost serving throughput by up to 89.2% while preserving accuracy.
-
Sparse Attention Remapping with Clustering for Efficient LLM Decoding on PIM
STARC remaps sparse KV caches by semantic clustering for PIM hardware, delivering 19-31% lower attention latency and 19-27% lower energy versus token-wise sparsity, with larger gains under tight KV budgets.