Generalization in LLM Problem Solving: The Case of the Shortest Path

• URL: http://arxiv.org/abs/2604.15306v1
• Date: Thu, 16 Apr 2026 17:59:43 GMT
• Title: Generalization in LLM Problem Solving: The Case of the Shortest Path
• Abstract summary: We introduce a controlled synthetic environment based on shortest-path planning.<n>We find that models exhibit strong spatial transfer but consistently fail under length scaling.<n>We analyze how distinct stages of the learning pipeline influence systematic problem-solving.
• Score: 20.04578544593786
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Whether language models can systematically generalize remains actively debated. Yet empirical performance is jointly shaped by multiple factors such as training data, training paradigms, and inference-time strategies, making failures difficult to interpret. We introduce a controlled synthetic environment based on shortest-path planning, a canonical composable sequential optimization problem. The setup enables clean separation of these factors and supports two orthogonal axes of generalization: spatial transfer to unseen maps and length scaling to longer-horizon problems. We find that models exhibit strong spatial transfer but consistently fail under length scaling due to recursive instability. We further analyze how distinct stages of the learning pipeline influence systematic problem-solving: for example, data coverage sets capability limits; reinforcement learning improves training stability but does not expand those limits; and inference-time scaling enhances performance but cannot rescue length-scaling failures.

Related papers

• How LLMs Fail and Generalize in RTL Coding for Hardware Design? [56.361436215029045]
  We introduce a new error taxonomy grounded in problem solvability, inspired by cognitive theory.<n>Our taxonomy categorizes failures into syntactic, semantic, solvable functional, and unsolvable functional types.
  arXiv  Detail & Related papers  (2026-04-26T14:34:49Z)
• Compensating Visual Insufficiency with Stratified Language Guidance for Long-Tail Class Incremental Learning [50.795452298625996]
  Long-tail class incremental learning (LT CIL) remains highly challenging because the scarcity of samples in tail classes hampers their learning.<n>We exploit the informativeness and scalability of language knowledge to tackle these issues.<n>We introduce stratified adaptive language guidance, which leverages learnable weights to merge multi-scale semantic representations.
  arXiv  Detail & Related papers  (2026-03-23T08:50:00Z)
• Bidirectional Curriculum Generation: A Multi-Agent Framework for Data-Efficient Mathematical Reasoning [16.95900718416944]
  We introduce a novel Bidirectional Curriculum Generation framework to maximize the instructional value of every training sample.<n>Unlike rigid trajectories, our multi-agent ecosystem mimics adaptive pedagogy to establish a closed feedback loop.<n>This mechanism ensures that the model consumes only the most effective data at any given stage.
  arXiv  Detail & Related papers  (2026-03-05T12:49:21Z)
• Provable Benefit of Curriculum in Transformer Tree-Reasoning Post-Training [76.12556589212666]
  We show that curriculum post-training avoids the exponential complexity bottleneck.<n>Under outcome-only reward signals, reinforcement learning finetuning achieves high accuracy with sample complexity.<n>We establish guarantees for test-time scaling, where curriculum-aware querying reduces both reward oracle calls and sampling cost from exponential to order.
  arXiv  Detail & Related papers  (2025-11-10T18:29:54Z)
• Structural Information-based Hierarchical Diffusion for Offline Reinforcement Learning [13.839214658191038]
  We propose Structural Information-based Hierarchical Diffusion framework for effective and stable offline policy learning.<n>We analyze structural information embedded in offline trajectories to construct the diffusion hierarchy adaptively.<n>We show that SIHD significantly outperforms state-of-the-art baselines in decision-making performance.
  arXiv  Detail & Related papers  (2025-09-26T06:24:06Z)
• Text-Trained LLMs Can Zero-Shot Extrapolate PDE Dynamics [10.472535430038759]
  Large language models (LLMs) have demonstrated emergent in-context learning (ICL) capabilities across a range of tasks.<n>We show that text-trained foundation models can accurately predict dynamics from discretized partial differential equation (PDE) solutions.<n>We analyze token-level output distributions and uncover a consistent ICL progression: beginning with syntactic pattern imitation, transitioning through an exploratory high-entropy phase, and culminating in confident, numerically grounded predictions.
  arXiv  Detail & Related papers  (2025-09-08T04:08:50Z)
• Data-Driven Exploration for a Class of Continuous-Time Indefinite Linear--Quadratic Reinforcement Learning Problems [6.859965454961918]
  We study reinforcement learning for continuous-time linear-quadratic (LQ) control problems.<n>We propose a model-free, data-driven exploration mechanism that adaptively adjusts entropy regularization by the critic.<n>Our method achieves a sublinear regret bound that matches the best-known model-free results for this class of LQ problems.
  arXiv  Detail & Related papers  (2025-07-01T01:09:06Z)
• RL for Reasoning by Adaptively Revealing Rationales [36.50924054394857]
  Supervised fine-tuning (SFT) relies on dense ground-truth labels, which become increasingly costly as sequence length grows.<n>We address this by adaptive backtracking (AdaBack), a per-sample curriculum learning algorithm that reveals only a partial prefix of the target output during training.<n>We show that our adaptive curriculum over partial answers reliably solves problems that are otherwise intractable.
  arXiv  Detail & Related papers  (2025-06-22T17:46:14Z)
• Growing Through Experience: Scaling Episodic Grounding in Language Models [67.27024505353384]
  Language models (LMs) require robust episodic grounding to excel at physical planning tasks.<n>Current episodic grounding approaches struggle with scalability and integration.<n>We propose a scalable weak-to-strong episodic learning framework that effectively transfers episodic behaviors from smaller to larger LMs.
  arXiv  Detail & Related papers  (2025-06-02T04:52:19Z)
• LESA: Learnable LLM Layer Scaling-Up [57.0510934286449]
  Training Large Language Models (LLMs) from scratch requires immense computational resources, making it prohibitively expensive.<n>Model scaling-up offers a promising solution by leveraging the parameters of smaller models to create larger ones.<n>We propose textbfLESA, a novel learnable method for depth scaling-up.
  arXiv  Detail & Related papers  (2025-02-19T14:58:48Z)
• SHERL: Synthesizing High Accuracy and Efficient Memory for Resource-Limited Transfer Learning [63.93193829913252]
  We propose an innovative METL strategy called SHERL for resource-limited scenarios. In the early route, intermediate outputs are consolidated via an anti-redundancy operation. In the late route, utilizing minimal late pre-trained layers could alleviate the peak demand on memory overhead.
  arXiv  Detail & Related papers  (2024-07-10T10:22:35Z)
• Provable Reinforcement Learning with a Short-Term Memory [68.00677878812908]
  We study a new subclass of POMDPs, whose latent states can be decoded by the most recent history of a short length $m$. In particular, in the rich-observation setting, we develop new algorithms using a novel "moment matching" approach with a sample complexity that scales exponentially. Our results show that a short-term memory suffices for reinforcement learning in these environments.
  arXiv  Detail & Related papers  (2022-02-08T16:39:57Z)

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.