Skip to content

jvdillon/sic

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

69 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Speed is Confidence

This repository contains the "Speed is Confidence" research paper and its associated experiments.

Abstract

Biological neural systems must be fast but are energy-constrained. Evolution's solution: act on the first signal. Winner-take-all circuits and time-to-first-spike coding implicitly treat when a neuron fires as an expression of confidence.

We apply this principle to ensembles of Tiny Recursive Models (TRM) [Jolicoeur-Martineau et al., 2025]. On Sudoku-Extreme, halt-first selection achieves 97% accuracy vs. 91% for probability averaging--while requiring 10x fewer reasoning steps. A single baseline model achieves 85.5% +/- 1.3%.

Can we internalize this as a training-only cost? Yes: by maintaining K=4 parallel latent states but backpropping only through the lowest-loss "winner," we achieve 96.9% +/- 0.6% accuracy--matching ensemble performance at 1x inference cost, with less than half the variance of the baseline. A key diagnostic: 89% of baseline failures are selection problems, revealing a 99% accuracy ceiling.

As in nature, this work was also resource constrained: all experiments used a single RTX 5090. A modified SwiGLU [Shazeer, 2020] made Muon [Jordan et al., 2024] and high LR viable, enabling baseline training in 48 minutes and full WTA (K=4) in 6 hours--compared to TRM's 20 hours on an L40S.

Installation

sudo apt install curl -y
curl -LsSf https://astral.sh/uv/install.sh | sh
uv sync
uv run code/sudoku/x182.py  # Train K=4, for example.

Much the code here builds on the excellent work of Alexia Jolicoeur-Martineau available from TinyRecursiveModels.

Reference

If you find our work useful, please consider citing:

@misc{dillon2026speedisconfidence,
      title={Speed is Confidence},
      author={Joshua V. Dillon},
      year={2026},
      eprint={2601.19085},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2601.19085},
}

and the Tiny Recursive Models paper:

@misc{jolicoeurmartineau2025morerecursivereasoningtiny,
      title={Less is More: Recursive Reasoning with Tiny Networks},
      author={Alexia Jolicoeur-Martineau},
      year={2025},
      eprint={2510.04871},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2510.04871},
}

License

Apache-2.0

About

Speed is Confidence

Resources

License

Stars

Watchers

Forks

Packages

No packages published