Saved in:
Bibliographic Details
Main Authors: Ren, Jiyuan, Du, Zhaocheng, Wen, Zhihao, Jia, Qinglin, Dai, Sunhao, Wu, Chuhan, Dong, Zhenhua
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2502.08661
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929717109587968
author Ren, Jiyuan
Du, Zhaocheng
Wen, Zhihao
Jia, Qinglin
Dai, Sunhao
Wu, Chuhan
Dong, Zhenhua
author_facet Ren, Jiyuan
Du, Zhaocheng
Wen, Zhihao
Jia, Qinglin
Dai, Sunhao
Wu, Chuhan
Dong, Zhenhua
contents As large language models (LLMs) advance, their ability to perform in-context learning and few-shot language generation has improved significantly. This has spurred using LLMs to produce high-quality synthetic data to enhance the performance of smaller models like online retrievers or weak LLMs. However, LLM-generated synthetic data often differs from the real data in key language attributes (e.g., styles, tones, content proportions, etc.). As a result, mixing these synthetic data directly with real data may distort the original data distribution, potentially hindering performance improvements. To solve this, we introduce SynAlign: a synthetic data generation and filtering framework based on key attribute distribution matching. Before generation, SynAlign employs an uncertainty tracker surrogated by the Gaussian Process model to iteratively select data clusters distinct from selected ones as demonstrations for new data synthesis, facilitating the efficient exploration diversity of the real data. Then, a latent attribute reasoning method is employed: the LLM summarizes linguistic attributes of demonstrations and then synthesizes new data based on them. This approach facilitates synthesizing diverse data with linguistic attributes that appear in real data.After generation, the Maximum Mean Discrepancy is used as the objective function to learn the sampling weight of each synthetic data, ensuring distribution matching with the real data. Our experiments on multiple text prediction tasks show significant performance improvements. We also conducted an online A/B test on an online retriever to demonstrate SynAlign's effectiveness.
format Preprint
id arxiv_https___arxiv_org_abs_2502_08661
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Few-shot LLM Synthetic Data with Distribution Matching
Ren, Jiyuan
Du, Zhaocheng
Wen, Zhihao
Jia, Qinglin
Dai, Sunhao
Wu, Chuhan
Dong, Zhenhua
Computation and Language
Artificial Intelligence
As large language models (LLMs) advance, their ability to perform in-context learning and few-shot language generation has improved significantly. This has spurred using LLMs to produce high-quality synthetic data to enhance the performance of smaller models like online retrievers or weak LLMs. However, LLM-generated synthetic data often differs from the real data in key language attributes (e.g., styles, tones, content proportions, etc.). As a result, mixing these synthetic data directly with real data may distort the original data distribution, potentially hindering performance improvements. To solve this, we introduce SynAlign: a synthetic data generation and filtering framework based on key attribute distribution matching. Before generation, SynAlign employs an uncertainty tracker surrogated by the Gaussian Process model to iteratively select data clusters distinct from selected ones as demonstrations for new data synthesis, facilitating the efficient exploration diversity of the real data. Then, a latent attribute reasoning method is employed: the LLM summarizes linguistic attributes of demonstrations and then synthesizes new data based on them. This approach facilitates synthesizing diverse data with linguistic attributes that appear in real data.After generation, the Maximum Mean Discrepancy is used as the objective function to learn the sampling weight of each synthetic data, ensuring distribution matching with the real data. Our experiments on multiple text prediction tasks show significant performance improvements. We also conducted an online A/B test on an online retriever to demonstrate SynAlign's effectiveness.
title Few-shot LLM Synthetic Data with Distribution Matching
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2502.08661