Shall Your Data Strategy Work? Perform a Swift Study

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Peng, Minlong, Yang, Jingyi, He, Zhongjun, Wu, Hua
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913698394669056
author Peng, Minlong
Yang, Jingyi
He, Zhongjun
Wu, Hua
author_facet Peng, Minlong
Yang, Jingyi
He, Zhongjun
Wu, Hua
contents This work presents a swift method to assess the efficacy of particular types of instruction-tuning data, utilizing just a handful of probe examples and eliminating the need for model retraining. This method employs the idea of gradient-based data influence estimation, analyzing the gradient projections of probe examples from the chosen strategy onto evaluation examples to assess its advantages. Building upon this method, we conducted three swift studies to investigate the potential of Chain-of-thought (CoT) data, query clarification data, and response evaluation data in enhancing model generalization. Subsequently, we embarked on a validation study to corroborate the findings of these swift studies. In this validation study, we developed training datasets tailored to each studied strategy and compared model performance with and without the use of these datasets. The results of the validation study aligned with the findings of the swift studies, validating the efficacy of our proposed method.
format Preprint
id arxiv_https___arxiv_org_abs_2502_13514
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Shall Your Data Strategy Work? Perform a Swift Study
Peng, Minlong
Yang, Jingyi
He, Zhongjun
Wu, Hua
Computation and Language
This work presents a swift method to assess the efficacy of particular types of instruction-tuning data, utilizing just a handful of probe examples and eliminating the need for model retraining. This method employs the idea of gradient-based data influence estimation, analyzing the gradient projections of probe examples from the chosen strategy onto evaluation examples to assess its advantages. Building upon this method, we conducted three swift studies to investigate the potential of Chain-of-thought (CoT) data, query clarification data, and response evaluation data in enhancing model generalization. Subsequently, we embarked on a validation study to corroborate the findings of these swift studies. In this validation study, we developed training datasets tailored to each studied strategy and compared model performance with and without the use of these datasets. The results of the validation study aligned with the findings of the swift studies, validating the efficacy of our proposed method.
title Shall Your Data Strategy Work? Perform a Swift Study
topic Computation and Language
url https://arxiv.org/abs/2502.13514