paper-with-me

홈 › Papers

Pencil Puzzle Bench: A Benchmark for Multi-Step Verifiable Reasoning

2026-03-02 · Justin Waugh arxiv

We introduce Pencil Puzzle Bench, a framework for evaluating large language model reasoning through pencil puzzles, a family of constraint-satisfaction problems closely related to NP-complete problems, with deterministic, step-level verification. From a database of 62,231 puzzles across 94 varieties with verified unique solutions, we select a benchmark of 300 puzzles spanning 20 varieties and evaluate 51 models from 11 providers in two modes: direct ask (single-shot) and agentic (multi-turn with iterative verification). A key differentiator of our benchmark is that every intermediate board state can be checked against variety-specific constraints, localizing errors to the exact rule violated, providing the infrastructure for dense, per-move reward signals for process supervision and reinforcement learning. Our evaluation reveals two distinct axes of capability: (1) reasoning effort scaling, where GPT-5.2 improves 81x from no reasoning to maximum effort; and (2) agentic iteration, where Claude Opus 4.6 rises from 0.3% to 30.0% through iterative checking, while GPT-5.2@xhigh improves from 20.2% to 56.0%. Agentic attempts span a median of 29 turns over 17 minutes, with the longest exceeding 1,221 turns and 14.3 hours - a demanding test of long-context utilization, not just reasoning.

📄 PDF Abstract BibTeX arXiv:2603.02119

Code (0)

등록된 구현이 없습니다.

Tasks

Reinforcement Learning

Similar Papers 제목 키워드 기반

Mathematical Definition and Systematization of Puzzle Rules

2024-12-18 · Itsuki Maeda, Yasuhiro Inoue

While logic puzzles have engaged individuals through problem-solving and critical thinking, the creation of new puzzle rules has largely relied on ad-hoc processes. Pencil puzzles, such as Slitherlink and Sudoku, represe…

Game DesignSpatial Reasoning

Probabilistic Tiny Recursive Model

2026-05-19 · Amin Sghaier, Ali Parviz, Alexia Jolicoeur-Martineau arxiv

Tiny Recursive Models (TRM) solve complex reasoning tasks with a fraction of the parameters of modern large language models (LLMs) by iteratively refining a latent state and final answer. While powerful, their determinis…

PENCIL: Long Thoughts with Short Memory

2025-03-18 · Chenxiao Yang, Nathan Srebro, David Mcallester, Zhiyuan Li

While recent works (e.g. o1, DeepSeek R1) have demonstrated great promise of using long Chain-of-Thought (CoT) to improve reasoning capabilities of language models, scaling it up during test-time is challenging due to in…

PRISM-Bench: A Benchmark of Puzzle-Based Visual Tasks with CoT Error Detection

2025-10-27 · Yusu Qian, Cheng Wan, Chao Jia, Yinfei Yang 외 arxiv

Multimodal large language models (MLLMs) have achieved remarkable progress on vision-language tasks, yet their reasoning processes remain sometimes unreliable. We introduce PRISM-Bench, a benchmark of puzzle-based visual…

Multimodal ReasoningAnswer GenerationVisual Reasoning

EnigmaEval: A Benchmark of Long Multimodal Reasoning Challenges

2025-02-13 · Clinton J. Wang, Dean Lee, Cristina Menghini, Johannes Mols 외

As language models master existing reasoning benchmarks, we need new challenges to evaluate their cognitive frontiers. Puzzle-solving events are rich repositories of challenging multimodal problems that test a wide range…

Humanity's Last ExamMultimodal Reasoning