CLAIMay 23, 2024

JiuZhang3.0: Efficiently Improving Mathematical Reasoning by Training Small Data Synthesis Models

arXiv:2405.14365v155 citationsh-index: 25Has CodeNIPS
Originality Incremental advance
AI Analysis

This reduces costs for researchers and practitioners needing efficient math reasoning models, but it is incremental as it builds on existing synthesis and distillation methods.

The paper tackles the high cost of enhancing mathematical reasoning in large language models by training a small LLM to synthesize math problems, reducing GPT-4 API calls to 9.3k and pre-training on 4.6B data, resulting in state-of-the-art performance on multiple datasets.

Mathematical reasoning is an important capability of large language models~(LLMs) for real-world applications. To enhance this capability, existing work either collects large-scale math-related texts for pre-training, or relies on stronger LLMs (\eg GPT-4) to synthesize massive math problems. Both types of work generally lead to large costs in training or synthesis. To reduce the cost, based on open-source available texts, we propose an efficient way that trains a small LLM for math problem synthesis, to efficiently generate sufficient high-quality pre-training data. To achieve it, we create a dataset using GPT-4 to distill its data synthesis capability into the small LLM. Concretely, we craft a set of prompts based on human education stages to guide GPT-4, to synthesize problems covering diverse math knowledge and difficulty levels. Besides, we adopt the gradient-based influence estimation method to select the most valuable math-related texts. The both are fed into GPT-4 for creating the knowledge distillation dataset to train the small LLM. We leverage it to synthesize 6 million math problems for pre-training our JiuZhang3.0 model, which only needs to invoke GPT-4 API 9.3k times and pre-train on 4.6B data. Experimental results have shown that JiuZhang3.0 achieves state-of-the-art performance on several mathematical reasoning datasets, under both natural language reasoning and tool manipulation settings. Our code and data will be publicly released in \url{https://github.com/RUCAIBox/JiuZhang3.0}.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes