paper
referenced-only
2023
paper:hanna-how-does-gpt-2-compute-greater-than-inte-2023How does GPT-2 compute greater-than?: Interpreting mathematical abilities in a pre-trained language model
Related work— refs + corpus + external arXiv
Cited / in-corpus / arXiv badges show which signals surfaced each row. Multi-source rows weighted higher.
- A Causal World Model Underlying Next Token Prediction: Exploring GPT in a Controlled EnvironmentYaniv Gurwicz, Sungduk Yu, Estelle Aflalo, Vasudev Lal Raanan Y. Rohekar2025≈ 74%
- How does GPT-2 Predict Acronyms? Extracting and Understanding a Circuit via Mechanistic InterpretabilityAlejandro Mat\'e, Juan Trujillo Jorge Garc\'ia-Carrasco2024≈ 73%
- LLMs achieve adult human performance on higher-order theory of mind tasksJohn Oliver Siy, Geoff Keeling, Adrien Baranes, Benjamin Barnett, Michael McKibben, Tatenda Kanyere, Alison Lentz, Blaise Aguera y Arcas, Robin I. M. Dunbar Winnie Street2024≈ 73%
- The use of GPT-4o and Other Large Language Models for the Improvement and Design of Self-Assessment Scales for Measurement of Interpersonal Communication SkillsGoran Buba\v{s}2024≈ 72%
- Interpretability and Generalization Bounds for Learning Spatial PhysicsTheo Gutman-Solo, Shuai Jiang Alejandro Francisco Queiruga2026≈ 72%
- What's Next in Affective Modeling? Large Language ModelsTobias Thejll-Madsen, Stacy Marsella Nutchanon Yongsatianchot2023≈ 72%
- Mechanistic interpretability of large language models with applications to the financial services industryKhashayar Filom, and Arjun Ravi Kannan Ashkan Golgoon2024≈ 72%
- Across the Levels of Analysis: Explaining Predictive Processing in Humans Requires More Than Machine-Estimated ProbabilitiesSathvik Nair and Colin Phillips2026≈ 72%
- Reanalyzing L2 Preposition Learning with Bayesian Mixed Effects and a Pretrained Language ModelJakob Prange and Man Ho Ivy Wong2026≈ 72%
- A geometric relation of the error introduced by sampling a language model's output distribution to its internal stateAlbert F. Modenbach2026≈ 72%
- Specialization after Generalization: Towards Understanding Test-Time Training in Foundation ModelsPatrik Wolf, Alexander Shevchenko, Dennis J\"uni, Andreas Krause, Gil Kur Jonas H\"ubotter2026≈ 72%
- Language Models Are Greedy Reasoners: A Systematic Formal Analysis of Chain-of-ThoughtAbulhair Saparov and He He2023≈ 71%
- Combining Insights From Multiple Large Language Models Improves Diagnostic AccuracyVictor Shia, Eugene Chu, Benjamin Harack, Nathan Fu Gioele Barabucci2024≈ 71%
- ≈ 71%
- An Automated Survey of Generative Artificial Intelligence: Large Language Models, Architectures, Protocols, and Applications\'Alvaro L\'opez L\'opez Eduardo C. Garrido-Merch\'an2026≈ 71%
- Evaluating Language Model Character Traitsin corpus2024≈ 70%
- Simulators — LessWrongin corpus≈ 70%
- ≈ 69%
- ≈ 68%
- Why Larger Models Learn More: Effects of Capacity, Interference, and Rare-Task Retentionin corpus2026≈ 67%
- ≈ 67%
- ≈ 67%
- Reasoning Models Don't Always Say What They Thinkin corpus2025≈ 67%
- ≈ 67%
- Open Problems in Mechanistic Interpretabilityin corpus2025≈ 67%
- Active Inference, Curiosity and Insightin corpus2017≈ 66%
- ≈ 66%
- ≈ 66%
- ≈ 66%
- ≈ 66%
Similar preprints — Semantic Scholar
Cited by (2)
- When Models Manipulate Manifolds: The Geometry of a Counting Task
Claude 3.5 Haiku's linebreaking in fixed-width text is implemented by manipulating low-dimensional, rippled 1-D feature manifolds: character count, line width, characters remaining and next-token leng
- 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