ε-coresets for attention exist of size O(√d e^{ρ+o(ρ)}/ε) for unit-norm keys/values and queries of norm ≤ρ, nearly matching the Ω(√d e^ρ/ε) lower bound.
Pyramidkv: Dynamic kv cache compression based on pyramidal information funneling.Conference on Language Modeling, 2025
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DS 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Nearly Optimal Attention Coresets
ε-coresets for attention exist of size O(√d e^{ρ+o(ρ)}/ε) for unit-norm keys/values and queries of norm ≤ρ, nearly matching the Ω(√d e^ρ/ε) lower bound.