Variational Latent Reasoning Guided by Rendered Chain-of-Thought

Fanmeng Wang, Haotian Liu, Guojiang Zhao, Hongteng Xu, Zhifeng Gao

Track: long paper (up to 10 pages)

Keywords: Large Language Models, Latent Reasoning

TL;DR: We propose a new and insightful latent reasoning paradigm that models latent reasoning within the VAE framework and learns it guided by rendered CoT, significantly outperforming existing latent reasoning methods.

Abstract:

While Chain-of-Thought (CoT) significantly enhances the performance of Large Language Models (LLMs), explicit reasoning chains introduce substantial computational redundancy.
Recent latent reasoning methods attempt to mitigate this by compressing reasoning processes into the high-dimensional latent space, but suffer from severe performance degradation due to the lack of appropriate compression guidance.
In this study, we propose Rendered CoT-Guided variational Latent Reasoning (ReGuLaR), a simple yet novel learning paradigm resolving this issue.
Fundamentally, we formulate latent reasoning within the Variational Auto-Encoding (VAE) framework, sampling the current latent reasoning state from the posterior distribution conditioned on previous ones.
Specifically, when learning this variational latent reasoning model, we render explicit reasoning chains as images, from which we extract dense visual-semantic representations to regularize the posterior distribution, thereby achieving efficient compression with minimal information loss.
Extensive experiments demonstrate that ReGuLaR significantly outperforms existing latent reasoning methods across both computational efficiency and reasoning effectiveness, and even surpasses CoT through multi-modal reasoning, providing a new and insightful solution to latent reasoning.
Code is available at https://github.com/FanmengWang/ReGuLaR.

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: 7