Recurrent Reasoning on Symbolic Puzzles with Sequence Models

• URL: http://arxiv.org/abs/2606.15686v1
• Date: Sun, 19 Apr 2026 15:48:32 GMT
• Title: Recurrent Reasoning on Symbolic Puzzles with Sequence Models
• Abstract summary: We introduce RecurrReason, a difficulty-controlled benchmark of four recurrent logic puzzles.<n>We benchmark two Transformer families, an encoder-decoder model (T5-style) and a decoder-only model (GPT-2-style)<n>Fine-tuned pre-trained T5 achieves 97.27% validation and 81.00% OOD accuracy on Block World; all models score 0.00% on River Crossing under all conditions.
• Score: 3.900695166480356
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models often appear strong on symbolic and algorithmic tasks, yet this apparent strength can hide brittle behaviour when problems become longer, harder, or slightly out of distribution. A major limitation of current reasoning benchmarks is that many primarily test whether a model can produce a valid answer, while paying less attention to whether the solution is minimal, robust, and stable under controlled difficulty scaling. We introduce RecurrReason, a difficulty-controlled benchmark of four recurrent logic puzzles (Tower of Hanoi, River Crossing, Block World, and Checkers Jumping) with BFS-optimal trajectories and a single interpretable difficulty parameter $N \in \{1,\dots,10\}$, totalling 10{,}817 unique puzzles and 285{,}933 moves. We benchmark two Transformer families, an encoder-decoder model (T5-style) and a decoder-only model (GPT-2-style), under consistent data splits and evaluation criteria, training on $N{=}1$ to $7$ and evaluating on both held-out in-distribution instances and harder out-of-distribution instances at $N{=}8$ to $10$. Fine-tuned pre-trained T5 achieves 97.27\% validation and 81.00\% OOD accuracy on Block World; all models score 0.00\% on River Crossing under all conditions. Failure mode analysis reveals that architecture is a stronger determinant of success than scale. Pre-training transfers only to puzzles with locally structured transition functions. Our code and dataset will be open-sourced upon acceptance.

Related papers

• Flow Reasoning Models: Scaling Reasoning Through Iterative Self-Refinement [13.205772367296353]
  We introduce Reasoning Flow Models, a training and test-time-scaling framework for structured reasoning with flow models.<n>A correct answer is a stable fixed point of the denoising dynamics, returning to itself when re-noised and re-solved.<n>We train flow models with a self-conditioning channel and close it at inference, letting them refine their own past predictions.
  arXiv  Detail & Related papers  (2026-06-28T02:10:36Z)
• Reasoning Quality Emerges Early: Data Curation for Reasoning Models [58.56882815783977]
  We show that diverse and challenging reasoning examples can be identified using only the initial reasoning tokens.<n>Our method outperforms existing baselines by up to 1.7% while being 91% more token efficient.
  arXiv  Detail & Related papers  (2026-06-25T09:32:58Z)
• LPDS: Evaluating LLM Robustness Through Logic-Preserving Difficulty Scaling [24.184930329823114]
  We introduce logic-state difficulty (LPDS) to quantifies the difficulty of a problem variation.<n>We show that as difficulty increases, performance declines and errors in the models' reasoning are more pronounced.
  arXiv  Detail & Related papers  (2026-05-14T20:26:59Z)
• BeyondBench: Benchmark-Free Evaluation of Reasoning in Language Models [13.380359214677176]
  We introduce BeyondBench, an evaluation framework that avoids contamination from internet-scale training data.<n>Our framework covers 44 algorithmic tasks with a total of 117 variations, grouped into three difficulty levels.<n>We evaluate 101 language models, including 85 open-source and 16 closed-source models, spanning sizes from 0.5B to 141B parameters.
  arXiv  Detail & Related papers  (2025-09-29T02:49:01Z)
• Staying in the Sweet Spot: Responsive Reasoning Evolution via Capability-Adaptive Hint Scaffolding [59.60915947702282]
  Reinforcement learning with verifiable rewards (RLVR) has achieved remarkable success in enhancing the reasoning capabilities of large language models (LLMs)<n>Existing RLVR methods often suffer from exploration inefficiency due to mismatches between the training data's difficulty and the model's capability.<n>We propose SEELE, a novel supervision-aided RLVR framework that dynamically adjusts problem difficulty to stay within the high-efficiency region.
  arXiv  Detail & Related papers  (2025-09-08T17:36:21Z)
• PuzzleClone: An SMT-Powered Framework for Synthesizing Verifiable Data [7.63617172069015]
  We introduce PuzzleClone, a framework for verifiable data at scale using Satisfiabilityo synthesizing (SMT)<n>Our approach features three key innovations: (1) encoding seed puzzles into structured logical specifications, (2) generating scalable variants through systematic variable and constraint randomization, and (3) ensuring validity via a reproduction mechanism.
  arXiv  Detail & Related papers  (2025-08-21T02:36:16Z)
• Concise Reasoning, Big Gains: Pruning Long Reasoning Trace with Difficulty-Aware Prompting [28.537281448659634]
  We propose a difficulty-aware prompting (DAP) method to dynamically shorten reasoning traces without performance loss.<n>In experiments, a student model fine-tuned on just 100K of these difficulty-pruned CoT samples outperforms a model distilled on 800K original Long CoT samples.<n>Our method also generalizes well: across 11 diverse benchmarks, the shorter difficulty-aware CoTs achieve equal or better accuracy than Long chains, using far fewer tokens.
  arXiv  Detail & Related papers  (2025-05-26T09:04:44Z)
• Climbing the Ladder of Reasoning: What LLMs Can-and Still Can't-Solve after SFT? [59.418994222096885]
  We conduct a detailed analysis of model performance on the AIME24 dataset.<n>We categorize questions into four tiers (Easy, Medium, Hard, and Extremely Hard)<n>We find that progression from Easy to Medium tier requires adopting an R1 reasoning style with minimal SFT-1K instances.<n>Exh-level questions present a fundamentally different challenge; they require unconventional problem-solving skills.
  arXiv  Detail & Related papers  (2025-04-16T03:39:38Z)
• LLMs Can Easily Learn to Reason from Demonstrations Structure, not content, is what matters! [53.84130385074551]
  Large reasoning models (LRMs) tackle complex reasoning problems by following long chain-of-thoughts (Long CoT)<n>We find that a Large Language model (LLM) can effectively learn Long CoT reasoning through data-efficient supervised fine-tuning (SFT) and parameter-efficient low-rank adaptation (LoRA)<n>With just 17k long CoT training samples, the Qwen2.5-32B-Instruct model achieves significant improvements on a wide range of math and coding benchmarks.
  arXiv  Detail & Related papers  (2025-02-11T08:48:48Z)
• Scaling up Test-Time Compute with Latent Reasoning: A Recurrent Depth Approach [70.44265766483633]
  We study a novel language model architecture that is capable of scaling test-time computation by implicitly reasoning in latent space.<n>Our model works by iterating a recurrent block, thereby unrolling to arbitrary depth at test-time.<n>We show that the resulting model can improve its performance on reasoning benchmarks, sometimes dramatically.
  arXiv  Detail & Related papers  (2025-02-07T18:55:02Z)
• s1: Simple test-time scaling [148.4204982041058]
  Test-time scaling is a promising new approach to language modeling that uses extra test-time compute to improve performance.<n>We seek the simplest approach to achieve test-time scaling and strong reasoning performance.
  arXiv  Detail & Related papers  (2025-01-31T18:48:08Z)
• DCR: Divide-and-Conquer Reasoning for Multi-choice Question Answering with LLMs [9.561022942046279]
  We propose Divide and Conquer Reasoning (DCR) to enhance the reasoning capability of large language models (LLMs) We first categorize questions into two subsets based on confidence score ($mathcalCS$), which is estimated by statistical frequency of generated answers. In particular, we first categorize questions into two subsets based on confidence score ($mathcalCS$), which is estimated by statistical frequency of generated answers.
  arXiv  Detail & Related papers  (2024-01-10T14:38:46Z)

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.