@YouJiacheng: > Directly applying RoPE rotation to KV will leak positional information into value matrix V 科学空间亦有记载 https://kexue.fm/…

X AI KOLs Timeline News

Summary

A social media post discusses the technical implication of applying RoPE rotation directly to KV caches, noting that it leaks positional information into the value matrix V.

> Directly applying RoPE rotation to KV will leak positional information into value matrix V 科学空间亦有记载 https://t.co/fn6SAS6fnk https://t.co/1AnXz4PTch
Original Article
View Cached Full Text

Cached at: 05/08/26, 09:53 AM

> Directly applying RoPE rotation to KV will leak positional information into value matrix V

科学空间亦有记载 https://t.co/fn6SAS6fnk https://t.co/1AnXz4PTch

Similar Articles

RoPE-Aware Bit Allocation for KV-Cache Quantization

arXiv cs.LG

Proposes Block-GTQ, a RoPE-aware bit allocation method for key-value cache quantization that improves long-context performance and memory efficiency by allocating more bits to high-energy RoPE blocks.

RoVE: Rotary Value Embeddings Attention for Relative Position-dependent Value Pathways

arXiv cs.LG

This paper proposes RoVE, a parameter-free modification to Rotary Position Embeddings that makes value pathways position-sensitive by rotating values simultaneously with keys, transforming RoPE attention into attentive convolution. Experiments on GPT-2 models show consistent gains in few-shot in-context learning, out-of-distribution perplexity, and long-context retrieval.