One Step Forward and K Steps Back: Better Reasoning with Denoising Recursion Models

Chris Cameron, Wangzheng Wang, Nikita Ivanov, Ashmita Bhattacharyya, Didier Chételat, Yingxue Zhang

Track: long paper (up to 10 pages)

Keywords: TRM, looped transformer, iterative refinement, ARC-AGI, latent reasoning, few-shot learning

TL;DR: We introduce Denoising Recursion Models, a method that corrupts data with noise like diffusion and trains the model to reverse it over multiple recursive steps like looped transformers.

Abstract: Looped transformers scale computational depth without increasing parameter count by repeatedly applying the same layer. However, training these models over long horizons creates significant optimization challenges. Specifically, it is difficult for looped transformers that start from noise to steer towards a complex output without additional supervision. Diffusion models tackle this issue by corrupting data with varying magnitudes of noise and training the model to reverse it in a single step. However, this process misaligns training and testing behaviour. We introduce Denoising Recursion Models, a method that similarly corrupts data with noise but trains the model to reverse the corruption over multiple recursive steps. This strategy provides a tractable curriculum of intermediate states, while better aligning training with testing and incentivizing non-greedy, forward-looking generation. Through extensive experiments, we show this approach outperforms the Tiny Recursion Model (TRM) on ARC-AGI, where it recently achieved breakthrough performance.

Anonymization: This submission has been anonymized for double-blind review via the removal of identifying information such as names, affiliations, and identifying URLs.

Submission Number: 77