Research map / Large language models / Agents, code and math
Solving Quantitative Reasoning Problems with Language Models
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| continued pretraining on technical/math web data | PaLM 540B · MATH | accuracy: 8.8 -> 33.6 | Helped |
| continued pretraining on technical/math web data | PaLM 8B · MATH | accuracy: 1.5 -> 14.1 | Helped |
| continued pretraining on technical/math web data | PaLM 540B · GSM8K | accuracy: 56.5 -> 58.8 | Helped |
| majority voting maj1@k instead of greedy decoding | Minerva 540B · MATH | accuracy: 33.6 -> 50.3 | Helped |
| majority voting maj1@k instead of greedy decoding | Minerva 540B · GSM8K | accuracy: 58.8 -> 78.5 | Helped |
| scale model from 62B to 540B | Minerva · MATH | accuracy: 27.6 -> 33.6 | Helped |
| majority voting instead of log-likelihood reranking (ablation) | Minerva · MATH | accuracy | Helped |
| finetune on MATH training set (ablation) | Minerva · MATH | accuracy | No clear effect |
Closest papers by meaning
- Physics of Language Models: Part 2.2, How to Learn From Mistakes on Grade-School Math Problems (ICLR 2025 · similarity 0.88)
- UGMathBench: A Diverse and Dynamic Benchmark for Undergraduate-Level Mathematical Reasoning with Large Language Models (ICLR 2025 · similarity 0.88)
- Exploring Equation as a Better Intermediate Meaning Representation for Numerical Reasoning of Large Language Models (AAAI 2024 · similarity 0.88)
- Advancing Mathematical Reasoning in Language Models: The Impact of Problem-Solving Data, Data Synthesis Methods, and Training Stages (ICLR 2025 · similarity 0.87)
- Physics of Language Models: Part 2.1, Grade-School Math and the Hidden Reasoning Process (ICLR 2025 · similarity 0.87)
- SciBench: Evaluating College-Level Scientific Problem-Solving Abilities of Large Language Models (ICML 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.