Glossary · Term

Parcae

← all terms

Definition

Plain language

A stable looped language model design used as a baseline for attractor-style architectures.

As stated in the literature

A 2026 looped-transformer language model architecture used as the primary baseline for Attractor Models on language modeling Pareto curves.

Why it matters: A clean, well-documented looped baseline makes claims about new iterative-reasoning architectures testable rather than vibes-based.

For example, Parcae uses fixed-point-style looping to read inputs, against which newer attractor architectures are compared on loss-versus-compute curves.

Heard on the show

“So on the language modeling side, the comparison they care about is against a model called Parcae — that's the current best stable looped language model, also a twenty-twenty-six paper.”
Episode 041 — When the Iteration Teaches the Model to Skip the Iteration

Mentioned in 1 episode

  1. 041
    When the Iteration Teaches the Model to Skip the Iteration

Related terms