Implicit SSMs, which iterate to a fixed point, combine RNN-style state tracking with parallel training and scale to 1.3B parameters.
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
CONDITIONAL 1representative citing papers
citing papers explorer
-
Implicit Language Models are RNNs: Balancing Parallelization and Expressivity
Implicit SSMs, which iterate to a fixed point, combine RNN-style state tracking with parallel training and scale to 1.3B parameters.