Shallow Queries, Mature Values: Depth-Asynchronous Self-Speculation for Looped Transformers.
arXiv:2609.34538v1 Announce Type: cross Abstract: Looped Transformers reuse a shared block across recurrent depths, making autoregressive decoding expensive because every generated token requires many sequential recurrent passes. Self-speculative decoders redu
arXiv cs.AIResearch source·