PDE-Controller: LLMs for Autoformalization and Reasoning of PDEs

M Soroco, J Song, M **a, K Emond, W Sun… - arxiv preprint arxiv …, 2025 - arxiv.org
While recent AI-for-math has made strides in pure mathematics, areas of applied
mathematics, particularly PDEs, remain underexplored despite their significant real-world …

Lower Bounds for Chain-of-Thought Reasoning in Hard-Attention Transformers

A Amiri, X Huang, M Rofin, M Hahn - arxiv preprint arxiv:2502.02393, 2025 - arxiv.org
Chain-of-thought reasoning and scratchpads have emerged as critical tools for enhancing
the computational capabilities of transformers. While theoretical results show that polynomial …

Mathematical Reasoning in Large Language Models: Assessing Logical and Arithmetic Errors across Wide Numerical Ranges

S Shrestha, M Kim, K Ross - arxiv preprint arxiv:2502.08680, 2025 - arxiv.org
Mathematical reasoning in Large Language Models (LLMs) is often evaluated using
benchmarks with limited numerical ranges, failing to reflect real-world problem-solving …

Low-Bit Quantization Favors Undertrained LLMs: Scaling Laws for Quantized LLMs with 100T Training Tokens

X Ouyang, T Ge, T Hartvigsen, Z Zhang, H Mi… - arxiv preprint arxiv …, 2024 - arxiv.org
We reveal that low-bit quantization favors undertrained large language models (LLMs) by
observing that models with larger sizes or fewer training tokens experience less quantization …

Fine Tuning Large Language Models to Deliver CBT for Depression

T Tahir - arxiv preprint arxiv:2412.00251, 2024 - arxiv.org
Cognitive Behavioral Therapy (CBT) is a well-established, evidence-based treatment for
Major Depressive Disorder. Unfortunately, there exist significant barriers to individuals …