---
title: "Latent Path Exploration via Potential-Guided Subtrajectory Balance"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2610.40283v1"
permalink: "https://feed.astralor.com/items/itm_672ea4b94113100e6043dbd1"
published: "2026-10-29T15:35:28.000Z"
score: 55
channel: "ai"
tags: ["RAG","多模态","Latent Path Exploration","Potential-Guided Subtrajectory Balance","Continuous","Advances","SubLAT","Subtrajectory Balance"]
---

# Latent Path Exploration via Potential-Guided Subtrajectory Balance

## At a glance

Continuous latent reasoning enables language models to perform intermediate computation without generating explicit reasoning traces. Advances in stochastic latent reasoning allow multiple paths to be sampled in continu…

## Key points

- Continuous latent reasoning enables language models to perform intermediate computation without generating explicit reasoning traces.
- Advances in stochastic latent reasoning allow multiple paths to be sampled in continuous space, but stochasticity alone does not ensure effective exploration.
- A key challenge is guiding stochastic transitions between intermediate latent states toward useful reasoning paths.

## Full text

Continuous latent reasoning enables language models to perform intermediate computation without generating explicit reasoning traces. Advances in stochastic latent reasoning allow multiple paths to be sampled in continuous space, but stochasticity alone does not ensure effective exploration. A key challenge is guiding stochastic transitions between intermediate latent states toward useful reasoning paths. We introduce SubLAT, which couples a relative answer potential with Subtrajectory Balance to learn stochastic latent exploration. The relative potential evaluates the answer support of sampled states against replacement states, with a correctness signal at the terminal state. Subtrajectory Balance relates potential differences to forward and backward transition densities across subtrajectories of different lengths. The model is jointly trained with answer supervision, CoT supervision, and representation distillation. Experiments on arithmetic and deductive reasoning show increasing solution coverage with larger sampling budgets across in-domain and out-of-domain benchmarks, as well as improvements on several structural OOD benchmarks. Trajectory analyses reveal progressive divergence among sampled latent paths and higher mean relative potentials for correct trajectories across arithmetic benchmarks.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_672ea4b94113100e6043dbd1