TARPO: Token-Wise Latent-Explicit Reasoning via Action-Routing Policy Optimization

• URL: http://arxiv.org/abs/2606.05859v1
• Date: Thu, 04 Jun 2026 08:30:53 GMT
• Title: TARPO: Token-Wise Latent-Explicit Reasoning via Action-Routing Policy Optimization
• Abstract summary: TARPO is a pure RL framework that switches between discrete token generation and continuous latent reasoning at each step.<n> TARPO consistently outperforms existing explicit and latent reasoning RL baselines across diverse benchmarks.
• Score: 18.977861031175756
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Latent reasoning has emerged as a promising alternative to discrete Chain-of-Thought (CoT) in large language models (LLMs), enabling more expressive reasoning by operating over continuous representations. However, the inherently deterministic nature of continuous representations limits policy exploration in reinforcement learning (RL). To address this, we propose TARPO (Token-Wise Latent-Explicit Reasoning via Action-Routing Policy Optimization), a pure RL framework that adaptively switches between discrete token generation and continuous latent reasoning at each step. TARPO introduces a lightweight action head router that observes the current hidden state and samples a routing decision from a binary mode-selection space, preserving the stochasticity of discrete token sampling from the vocabulary. The LLM backbone and router are jointly optimized end-to-end with a shared group-relative advantage signal. Extensive experiments across Qwen2.5 (from 1.5B to 7B) and Llama-3.1-8B backbones demonstrate that TARPO consistently outperforms existing explicit and latent reasoning RL baselines across diverse benchmarks. Further analysis shows that TARPO learns adaptive token-wise switching behaviors while maintaining stable training dynamics. Our code is available at https://github.com/NKU-LITI/TARPO-master.

Related papers

• Efficient and Trainable Language Model Test-Time Scaling via Local Branch Routing [65.59464371673623]
  Local Branch Routing (LBR) is a token-level test-time scaling framework.<n>It expands a small local lookahead tree, forwards all sampled branches through the language model, and uses a lightweight router to select the depth-1 subtree to commit.<n>LBR allows each token decision to use evidence beyond the root next-token distribution while avoiding full solution-level search.
  arXiv  Detail & Related papers  (2026-06-24T03:42:44Z)
• LASAR: Latent Adaptive Semantic Aligned Reasoning for Generative Recommendation [33.48046116606003]
  Latent reasoning has emerged as an effective paradigm in Large Language Models (LLMs)<n>We propose LASAR (Latent Adaptive Semantic Aligned Reasoning), an SFT-then-RL framework.<n> Experiments on three real-world datasets demonstrate that LASAR outperforms all baselines.
  arXiv  Detail & Related papers  (2026-05-11T08:52:53Z)
• LEPO: Latent Reasoning Policy Optimization for Large Language Models [11.032175358561162]
  We introduce controllableity into latent reasoning via Gumbel-Softmax.<n>We propose textbfunderlineLatent Rtextbfunderlineeasoning textbfunderlinePolicy textbfunderlineOptimization(textbfLEPO)<n>In tests, LEPO significantly outperforms existing RL methods for discrete and latent reasoning.
  arXiv  Detail & Related papers  (2026-04-20T07:05:12Z)
• Bridging Perception and Reasoning: Token Reweighting for RLVR in Multimodal LLMs [51.60575965819268]
  We propose a plug-and-play Token-Reweighting (ToR) strategy that explicitly models this interdependence.<n>ToR delivers consistent performance gains across multiple multi-modal reasoning benchmarks.
  arXiv  Detail & Related papers  (2026-03-26T06:25:27Z)
• Rethinking Token-Level Policy Optimization for Multimodal Chain-of-Thought [73.39221516441624]
  Multimodal Chain-of-Thought (CoT) reasoning requires large vision-language models to construct reasoning trajectories.<n>Existing Reinforcement Learning with Verifiable Rewards (RLVR) methods treat CoT uniformly without distinguishing varying degrees of visual grounding.<n>We propose Perception-Exploration Policy Optimization (PEPO), which derives a perception prior to hidden state similarity and integrates it with token entropy.
  arXiv  Detail & Related papers  (2026-03-24T06:38:00Z)
• Lookahead Tree-Based Rollouts for Enhanced Trajectory-Level Exploration in Reinforcement Learning with Verifiable Rewards [48.321707628011005]
  Lookahead Tree-Based Rollouts (LATR) is a novel rollout strategy designed to explicitly promote trajectory-level diversity.<n>LATR accelerates policy learning by 131% on average and improves final pass@1 performance by 4.2%.
  arXiv  Detail & Related papers  (2025-10-28T11:12:02Z)
• Attention Illuminates LLM Reasoning: The Preplan-and-Anchor Rhythm Enables Fine-Grained Policy Optimization [56.083511902353365]
  Reinforcement learning (RL) typically applies uniform credit across an entire generation of Large language models.<n>This work positions attention as a privileged substrate that renders the internal logic of LLMs as a mechanistic blueprint of reasoning itself.<n>We introduce three novel RL strategies that dynamically perform targeted credit assignment to critical nodes.
  arXiv  Detail & Related papers  (2025-10-15T13:49:51Z)
• Parallel Test-Time Scaling for Latent Reasoning Models [58.428340345068214]
  Parallel test-time scaling (TTS) is a pivotal approach for enhancing large language models (LLMs)<n>Recent advances in latent reasoning, where intermediate reasoning unfolds in continuous vector spaces, offer a more efficient alternative to explicit Chain-of-Thought.<n>This work enables parallel TTS for latent reasoning models by addressing the above issues.
  arXiv  Detail & Related papers  (2025-10-09T03:33:00Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.