Imagen de Portada

The Illusion of Equivalence: Systematic FP16 Divergence in KV-Cached Autoregressive Inference

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Chodavarapu, Ranjith, Xu, Lei
Formato:	Preprint
Publicado:	2026
Materias:	Machine Learning Artificial Intelligence
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Uncertainty-Aware Wildfire Smoke Density Classification from Satellite Imagery via CBAM-Augmented EfficientNet with Evidential Deep Learning
por: Chodavarapu, Ranjith
Publicado: (2026)

Probabilistic Dating of Historical Manuscripts via Evidential Deep Regression on Visual Script Features
por: Chodavarapu, Ranjith
Publicado: (2026)

SmallKV: Small Model Assisted Compensation of KV Cache Compression for Efficient LLM Inference
por: Zhao, Yi, et al.
Publicado: (2025)

FreeKV: Boosting KV Cache Retrieval for Efficient LLM Inference
por: Liu, Guangda, et al.
Publicado: (2025)

Accelerating LLM Inference Throughput via Asynchronous KV Cache Prefetching
por: Dong, Yanhao, et al.
Publicado: (2025)

Online Scheduling for LLM Inference with KV Cache Constraints
por: Jaillet, Patrick, et al.
Publicado: (2025)

The Pitfalls of KV Cache Compression
por: Chen, Alex, et al.
Publicado: (2025)

SentenceKV: Efficient LLM Inference via Sentence-Level Semantic KV Caching
por: Zhu, Yuxuan, et al.
Publicado: (2025)

KeepKV: Achieving Periodic Lossless KV Cache Compression for Efficient LLM Inference
por: Tian, Yuxuan, et al.
Publicado: (2025)

The Residual Stream Is All You Need: On the Redundancy of the KV Cache in Transformer Inference
por: Qasim, Kaleem Ullah, et al.
Publicado: (2026)

Defeating the Training-Inference Mismatch via FP16
por: Qi, Penghui, et al.
Publicado: (2025)

Leyline: KV Cache Directives for Agentic Inference
por: Ma, Bole, et al.
Publicado: (2026)

KV Cache Transform Coding for Compact Storage in LLM Inference
por: Staniszewski, Konrad, et al.
Publicado: (2025)

CacheClip: Accelerating RAG with Effective KV Cache Reuse
por: Yang, Bin, et al.
Publicado: (2025)

ReasonCache: Accelerating Large Reasoning Model Serving through KV Cache Sharing
por: Chen, Kaiwen, et al.
Publicado: (2025)

Get More with LESS: Synthesizing Recurrence with KV Cache Compression for Efficient LLM Inference
por: Dong, Harry, et al.
Publicado: (2024)

KV Packet: Recomputation-Free Context-Independent KV Caching for LLMs
por: Chen, Chuangtao, et al.
Publicado: (2026)

CommonKV: Compressing KV Cache with Cross-layer Parameter Sharing
por: Wang, Yixuan, et al.
Publicado: (2025)

CoKV: Optimizing KV Cache Allocation via Cooperative Game
por: Sun, Qiheng, et al.
Publicado: (2025)

MadaKV: Adaptive Modality-Perception KV Cache Eviction for Efficient Multimodal Long-Context Inference
por: Li, Kunxi, et al.
Publicado: (2025)

TTKV: Temporal-Tiered KV Cache for Long-Context LLM Inference
por: Dzikanyanga, Gradwell, et al.
Publicado: (2026)

LouisKV: Efficient KV Cache Retrieval for Long Input-Output Sequences
por: Wu, Wenbo, et al.
Publicado: (2025)

ALISA: Accelerating Large Language Model Inference via Sparsity-Aware KV Caching
por: Zhao, Youpeng, et al.
Publicado: (2024)

RelayCaching: Accelerating LLM Collaboration via Decoding KV Cache Reuse
por: Geng, Yingsheng, et al.
Publicado: (2026)

ZipCache: Accurate and Efficient KV Cache Quantization with Salient Token Identification
por: He, Yefei, et al.
Publicado: (2024)

Cache What Lasts: Token Retention for Memory-Bounded KV Cache in LLMs
por: Bui, Ngoc, et al.
Publicado: (2025)

KV-Fold: One-Step KV-Cache Recurrence for Long-Context Inference
por: Nadali, Alireza, et al.
Publicado: (2026)

LookaheadKV: Fast and Accurate KV Cache Eviction by Glimpsing into the Future without Generation
por: Ahn, Jinwoo, et al.
Publicado: (2026)

KVSharer: Efficient Inference via Layer-Wise Dissimilar KV Cache Sharing
por: Yang, Yifei, et al.
Publicado: (2024)

Hurwitz Quaternion Multiplicative Quantization for KV Cache Compression
por: Swain, Kabir, et al.
Publicado: (2026)

Palu: Compressing KV-Cache with Low-Rank Projection
por: Chang, Chi-Chih, et al.
Publicado: (2024)

PolarQuant: Quantizing KV Caches with Polar Transformation
por: Han, Insu, et al.
Publicado: (2025)

Polynomial Context-Truncation Sensitivity in Autoregressive Language Models: Sequential Wyner-Ziv Bounds for KV Cache Compression
por: Kim, Munsik
Publicado: (2026)

ClusterKV: Manipulating LLM KV Cache in Semantic Space for Recallable Compression
por: Liu, Guangda, et al.
Publicado: (2024)

MARché: Fast Masked Autoregressive Image Generation with Cache-Aware Attention
por: Jiang, Chaoyi, et al.
Publicado: (2025)

GEAR: An Efficient KV Cache Compression Recipe for Near-Lossless Generative Inference of LLM
por: Kang, Hao, et al.
Publicado: (2024)

KVSculpt: KV Cache Compression as Distillation
por: Jiang, Bo, et al.
Publicado: (2026)

CONF-KV: Confidence-Aware KV Cache Eviction with Mixed-Precision Storage for Long-Horizon LLM
por: Li, Yubo, et al.
Publicado: (2026)

ReCalKV: Low-Rank KV Cache Compression via Head Reordering and Offline Calibration
por: Yan, Xianglong, et al.
Publicado: (2025)

AsymKV: Enabling 1-Bit Quantization of KV Cache with Layer-Wise Asymmetric Quantization Configurations
por: Tao, Qian, et al.
Publicado: (2024)