Latent PRM guidance selects among hidden-state trajectories to raise executable validation rate from 32.89% to 42.1% on a 76-task parallel code translation benchmark.
arXiv preprint arXiv:2411.08706 , year=
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
Denoising Recursion Models train multi-step noise reversal in looped transformers and outperform the prior Tiny Recursion Model on ARC-AGI.
citing papers explorer
-
Latent Reasoning Guidance for Parallel Code Translation
Latent PRM guidance selects among hidden-state trajectories to raise executable validation rate from 32.89% to 42.1% on a 76-task parallel code translation benchmark.
-
One Step Forward and K Steps Back: Better Reasoning with Denoising Recursion Models
Denoising Recursion Models train multi-step noise reversal in looped transformers and outperform the prior Tiny Recursion Model on ARC-AGI.