TConstFormer claims O(1) amortized attention via periodic recompression, but its own equations imply average per-token cost grows linearly with stream length because recompression recurs.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
From TLinFormer to TConstFormer: The Leap to Constant-Time Transformer Attention: Achieving O(1) Computation and O(1) KV Cache during Autoregressive Inference
TConstFormer claims O(1) amortized attention via periodic recompression, but its own equations imply average per-token cost grows linearly with stream length because recompression recurs.