Reproduce “Recurrent Looped Transformer” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Yifan Zhang, Jichen Feng, Shihan Qin, Sun-Yuan Kung, Mengdi Wang
RLT-1 gives a Transformer an updateable state across tokens by feeding each token’s final decoder output through a gated merge into the next token’s decoder input. This extends computation across a sequence while keeping blocks per token fixed, though attention work grows with context. On supervised parity tasks, it scores 100.00% at 256 bits versus GPT Layer 8’s 50.07%; generalization remains unestablished.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.