paper
referenced-only
2024
paper:y-arithmetic-without-algorithms-language-m-2024Arithmetic without algorithms: language models solve math with a bag of heuristics
ByNikankin Y·Reusch A·Mueller A·Belinkov Y
Related work— refs + corpus + external arXiv
Cited / in-corpus / arXiv badges show which signals surfaced each row. Multi-source rows weighted higher.
- Language Models Are Greedy Reasoners: A Systematic Formal Analysis of Chain-of-ThoughtAbulhair Saparov and He He2023≈ 74%
- A Pattern Language for Machine Learning TasksIan Fan, Tuomas Laakkonen, Neil John Ortega, Thomas Hoffmann, Vincent Wang-Mascianica Benjamin Rodatz2025≈ 74%
- Do Language Models Follow Occam's Razor? An Evaluation of Parsimony in Inductive and Abductive ReasoningAbulhair Saparov Yunxin Sun2026≈ 73%
- Reasoning with Language Model is Planning with World ModelYi Gu, Haodi Ma, Joshua Jiahua Hong, Zhen Wang, Daisy Zhe Wang, Zhiting Hu Shibo Hao2023≈ 73%
- ACES: Generating Diverse Programming Puzzles with with Autotelic Generative ModelsC\'edric Colas, Gaia Molinaro, Pierre-Yves Oudeyer, Laetitia Teodorescu Julien Pourcel2026≈ 73%
- Evaluating the World Model Implicit in a Generative ModelJustin Y. Chen, Ashesh Rambachan, Jon Kleinberg, Sendhil Mullainathan Keyon Vafa2024≈ 72%
- The Model Says Walk: How Surface Heuristics Override Implicit Constraints in LLM ReasoningLu Zhang, Tianchong Jiang, Ramayya Krishnan, Rema Padman Yubo Li2026≈ 72%
- What do Language Models Learn and When? The Implicit Curriculum HypothesisKaiser Sun, Millicent Li, Isabelle Lee, Lindia Tjuatja, Jen-tse Huang, Graham Neubig Emmy Liu2026≈ 72%
- Explanation through Reward Model Reconciliation using POMDP Tree SearchAnshu Saksena, Anna L. Buczak, Zachary N. Sunberg Benjamin D. Kraske2026≈ 71%
- Seemingly Simple Planning Problems are Computationally Challenging: The Countdown GameHarsha Kokel, Sarath Sreedharan Michael Katz2026≈ 71%
- Across the Levels of Analysis: Explaining Predictive Processing in Humans Requires More Than Machine-Estimated ProbabilitiesSathvik Nair and Colin Phillips2026≈ 71%
- ≈ 71%
- The Information-theoretic and Algorithmic Approach to Human, Animal and Artificial CognitionHector Zenil, Jesper Tegn\'er Nicolas Gauvrit2015≈ 71%
- Comprehension Without Competence: Architectural Limits of LLMs in Symbolic Computation and ReasoningZheng Zhang2025≈ 71%
- A geometric relation of the error introduced by sampling a language model's output distribution to its internal stateAlbert F. Modenbach2026≈ 71%
- ≈ 69%
- ≈ 69%
- Learning without neurons in physical systemsin corpus2022≈ 68%
- ≈ 68%
- Opening the Hood of a Word Processorin corpus1984≈ 67%
- Active Inference, Curiosity and Insightin corpus2017≈ 67%
- ≈ 66%
- ≈ 66%
- Linda in contextin corpus1989≈ 66%
- ≈ 66%
- ≈ 66%
- Active inference: demystified and comparedin corpus2021≈ 65%
- ≈ 65%
- The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasetsin corpus2023≈ 65%
Similar preprints — Semantic Scholar
Cited by (3)
- Testing the Limits of Truth Directions in LLMs
Linear truth directions in LLMs are reliable primarily for simple factual retrieval and break down as soon as truth assessment requires tracking intermediate results—a finding that sharply constrains
- Arithmetic in the Wild: Llama uses Base-10 Addition to Reason About Cyclic Concepts
Llama-3.1-8B solves cyclic arithmetic (e.g., "what month is six months after August?") not by performing modular addition in the period of the cyclic concept (12 for months, 7 for days of the week) as
- World models, artificial general intelligence and the hard problems of life–mind continuity: toward a unified understanding of natural and artificial intelligence