Reasoning and the "physics" of language models — page 2

• In-Context Algebra
  arxiv:2512.16902 · 2025-12-18 · text available
• Incorrect Baseline Evaluations Call into Question Recent LLM-RL Claims
  url:safe-lip-9a8.notion.site/incorrect-baseline-evaluations-call-into-question-recent-llm-rl-claims-2012f1fbf0ee8094ab8ded1953c15a37 · 2025-05-29 · metadata only
• Is the Reversal Curse a Binding Problem? Uncovering Limitations of Transformers from a Basic Generalization Failure
  arxiv:2504.01928 · 2025-04-02 · text available
• Knowledge Mechanisms in Large Language Models: A Survey and Perspective
  arxiv:2407.15017 · 2024-07-22 · text available
• Language is primarily a tool for communication rather than thought
  doi:10.1038/s41586-024-07522-w · 2024-06-01 · text available
• Language Models Are Capable of Metacognitive Monitoring and Control of Their Internal Activations
  arxiv:2505.13763 · 2025-05-19 · text available
• Language Models Compare Quantities Using Number-specific and Unit-specific Heuristics
  arxiv:2606.03982 · 2026-06-02 · text available
• Language Models Use Trigonometry to Do Addition
  arxiv:2502.00873 · 2025-02-02 · text available
• Latent Recurrent Thoughts: Recurrent Refinement of Proposed Latents for Reasoning with Frozen LLMs
  arxiv:2609.01117 · 2026-09-01 · text available
• Less is More: Recursive Reasoning with Tiny Networks
  arxiv:2510.04871 · 2025-10-06 · text available
• LIMO: Less is More for Reasoning
  arxiv:2502.03387 · 2025-02-05 · text available
• LiveMathematicianBench: A Live Benchmark for Mathematician-Level Reasoning with Proof Sketches
  arxiv:2604.01754 · 2026-04-02 · text available
• LLMs Can't Plan, But Can Help Planning in LLM-Modulo Frameworks
  arxiv:2402.01817 · 2024-02-02 · text available
• Modular Arithmetic: Language Models Solve Math Digit by Digit
  arxiv:2508.02513 · 2025-08-04 · text available
• Multimodal Chain-of-Thought Reasoning in Language Models
  arxiv:2302.00923 · 2023-02-02 · text available
• Physics of Language Models: Part 1, Learning Hierarchical Language Structures
  arxiv:2305.13673 · 2023-05-23 · text available
• Physics of Language Models: Part 2.1, Grade-School Math and the Hidden Reasoning Process
  arxiv:2407.20311 · 2024-07-29 · text available
• Physics of Language Models: Part 3.1, Knowledge Storage and Extraction
  arxiv:2309.14316 · 2023-09-25 · text available
• Physics of Language Models: Part 3.2, Knowledge Manipulation
  arxiv:2309.14402 · 2023-09-25 · text available
• Physics of Language Models: Part 3.3, Knowledge Capacity Scaling Laws
  arxiv:2404.05405 · 2024-04-08 · text available
• PlanBench: An Extensible Benchmark for Evaluating Large Language Models on Planning and Reasoning about Change
  openreview:wUU-7XTL5XO · 2022-06-21 · text available
• Position: LLMs can't jump
  openreview:klU4737opt · 2026-01-27 · text available
• Pre-trained Large Language Models Use Fourier Features to Compute Addition
  arxiv:2406.03445 · 2024-06-05 · text available
• Progress measures for grokking via mechanistic interpretability
  arxiv:2301.05217 · 2023-01-12 · text available