paper-with-me

Papers

How Far Ahead Do LLMs Plan? Uncovering the Latent Horizon in Chain-of-Thought Reasoning

2026-02-02 · Liyan Xu, Mo Yu, Fandong Meng, Jie Zhou arxiv

Chain-of-thought (CoT) reasoning has become a central mechanism for eliciting multi-step reasoning in Large Language Models (LLMs). Yet recent evidence presents a tension: hidden states appear to already encode future reasoning before CoT fully unfolds, while explicit steps still remain crucial for tasks requiring compositional computation. To deepen the understanding between LLM's internal states and its verbalized reasoning trajectories, we investigate the latent planning strength of LLMs, through our probing method, Tele-Lens, applying to hidden states across diverse task domains. Our empirical results indicate that LLMs exhibit a myopic horizon, primarily conducting incremental transitions without precise global planning. Leveraging this characteristic, we propose a hypothesis on enhancing uncertainty estimation of CoT, which we validate that a sparse set of pivot positions can effectively represent the uncertainty of the entire path. We further underscore the significance of exploiting CoT dynamics, and demonstrate that automatic recognition of CoT bypass can be achieved without performance degradation. Our code, data and models are released at https://github.com/lxucs/tele-lens.

📄 PDF Abstract BibTeX arXiv:2602.02103

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Latent Planning Emerges with Scale

2026-04-14 · Michael Hanna, Emmanuel Ameisen arxiv

LLMs can perform seemingly planning-intensive tasks, like writing coherent stories or functioning code, without explicitly verbalizing a plan; however, the extent to which they implicitly plan is unknown. In this paper, …

ELVIS: Ensemble-Calibrated Latent Imagination for Long-Horizon Visual MPC

2026-05-06 · Yurui Du, Pinhao Song, Yutong Hu, Renaud Detry arxiv

A central challenge of visual control with model-based reinforcement learning (RL) is reliable long-horizon planning: long rollouts with learned latent dynamics exhibit branching futures and multi-modal action-value dist…

Reinforcement Learning

Look Further Ahead: Testing the Limits of GPT-4 in Path Planning

2024-06-17 · Mohamed Aghzal, Erion Plaku, Ziyu Yao

Large Language Models (LLMs) have shown impressive capabilities across a wide variety of tasks. However, they still face challenges with long-horizon planning. To study this, we propose path planning tasks as a platform …

Navigate

Planning and Learning with Adaptive Lookahead

2022-01-28 · Aviv Rosenberg, Assaf Hallak, Shie Mannor, Gal Chechik 외

Some of the most powerful reinforcement learning frameworks use planning for action selection. Interestingly, their planning horizon is either fixed or determined arbitrarily by the state visitation history. Here, we exp…

The Objective Is the Bottleneck: Latent World Models Encode What Their Planners Cannot Use

2026-08-13 · Joyjeet Singh arxiv

Latent world models are judged by how well they predict, so when planning fails at long horizons the natural reading is that the predictor degrades. On a reproduction of LeWorldModel on TwoRoom we show the binding constr…