Unverified50% confidenceFactExact time
OAL (Outlier-Aware Attention Linearization) is a dimension-aware method for constructing quadratic attention approximations.
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedGated Linear Attention (GLA) reduces computational complexity from O(N²) to O(N) through a linearized attention mechanism.67% similarUnverified标准Attention计算的显存占用和带宽消耗均为O(N²)，FlashAttention将显存占用降至O(N)67% similarUnverifiedEMA (Efficient Multi-scale Attention) is a lightweight attention mechanism that enhances joint channel-spatial representation through cross-dimensional multi-scale feature interaction, with lower computational overhead than standard self-attention.67% similarUnverified滑动窗口注意力（SWA）由 Mistral AI 广泛采用，将注意力复杂度从 O(N²) 降至 O(N×W)65% similarUnverifiedThe standard attention mechanism has O(n²) memory overhead, requiring computation of an approximately 2000×2000 attention matrix when there are 2,000 rows.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1002303API
curl https://kongchang.com/api/v1/knowledge/claims/1002303MCP
get_claim(id=1002303)