A Survey on Latent Reasoning

TL;DR

This survey reviews the emerging field of latent reasoning in large language models, focusing on methods that perform multi-step inference within hidden states to improve reasoning without explicit token-level supervision.

Contribution

It provides a comprehensive overview of latent reasoning techniques, including hierarchical neural representations, activation recurrence, and advanced models like masked diffusion, to clarify the field's landscape and future directions.

Findings

Latent reasoning enables multi-step inference in continuous hidden states.

Hierarchical neural representations support complex transformations.

Advanced models like masked diffusion facilitate globally consistent reasoning.

Abstract

Large Language Models (LLMs) have demonstrated impressive reasoning capabilities, especially when guided by explicit chain-of-thought (CoT) reasoning that verbalizes intermediate steps. While CoT improves both interpretability and accuracy, its dependence on natural language reasoning limits the model's expressive bandwidth. Latent reasoning tackles this bottleneck by performing multi-step inference entirely in the model's continuous hidden state, eliminating token-level supervision. To advance latent reasoning research, this survey provides a comprehensive overview of the emerging field of latent reasoning. We begin by examining the foundational role of neural network layers as the computational substrate for reasoning, highlighting how hierarchical representations support complex transformations. Next, we explore diverse latent reasoning methodologies, including activation-based recurrence, hidden state propagation, and fine-tuning strategies that compress or internalize explicit reasoning traces. Finally, we discuss advanced paradigms such as infinite-depth latent reasoning via masked diffusion models, which enable globally consistent and reversible reasoning processes. By unifying these perspectives, we aim to clarify the conceptual landscape of latent reasoning and chart future directions for research at the frontier of LLM cognition. An associated GitHub repository collecting the latest papers and repos is available at: https://github.com/multimodal-art-projection/LatentCoT-Horizon/.