Quantifying Error Tolerance in Synthetic Data: An Atomic-level Operand vs. Operator Perturbation Study

2026-08-29

AI Summary

本文通过引入Atomic Tree Operation Modeling (ATOM)框架，解决了合成数据生成中误差容忍度量化分析不足的问题，区分了操作数和操作符扰动，优化了数据过滤策略。

Abstract

Synthetic data generation has become a cornerstone for advancing large language models. However, the lack of the quantitative analysis for error tolerance became a critical bottleneck. Consequently, current filtering strategies fluctuate between two extremes: they are either overly aggressive, risking the exclusion of potentially valuable samples, or overly permissive, failing to eliminate erroneous samples effectively. To bridge this gap, this paper introduces Atomic Tree Operation Modeling (ATOM), a framework that decomposes data into functional units ($f(x)\rightarrow y$). ATOM distinguishes benign Operand $x$ perturbations from fatal Operator $f$ perturbations. The former are needlessly discarded by aggressive filtering, while the latter slip through permissive filtering. Our experiments reveal a double dissociation: models are robust to operand perturbations but collapse under operator perturbations. By prioritizing operator over aggressive operand precision, our ATOM-synthesized data outperforms rigorous baselines (e.g., +3.1% gain over LIMA), suggesting that operator diversity matters more than operand precision. Our code is available at https://github.com/Lut-hub/ATOM.

Problem

• error tolerance
• synthetic data generation
• filtering strategies

Innovation

• Atomic Tree Operation Modeling (ATOM)
• operand perturbations
• operator perturbations
• error tolerance