Pith. sign in

BERT on a Data Diet: Finding Important Examples by Gradient-Based Pruning

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it
abstract

Current pre-trained language models rely on large datasets for achieving state-of-the-art performance. However, past research has shown that not all examples in a dataset are equally important during training. In fact, it is sometimes possible to prune a considerable fraction of the training set while maintaining the test performance. Established on standard vision benchmarks, two gradient-based scoring metrics for finding important examples are GraNd and its estimated version, EL2N. In this work, we employ these two metrics for the first time in NLP. We demonstrate that these metrics need to be computed after at least one epoch of fine-tuning and they are not reliable in early steps. Furthermore, we show that by pruning a small portion of the examples with the highest GraNd/EL2N scores, we can not only preserve the test accuracy, but also surpass it. This paper details adjustments and implementation choices which enable GraNd and EL2N to be applied to NLP.

fields

cs.DB 1

years

2025 1

verdicts

CONDITIONAL 1

representative citing papers

A Survey of LLM $\times$ DATA

cs.DB · 2025-05-24 · conditional · novelty 5.0

A comprehensive survey of the bidirectional links between LLMs and data management, organized as DATA4LLM and LLM4DATA with a new 'IaaS' data-quality framework.

citing papers explorer

Showing 1 of 1 citing paper.

  • A Survey of LLM $\times$ DATA cs.DB · 2025-05-24 · conditional · none · ref 137 · internal anchor

    A comprehensive survey of the bidirectional links between LLMs and data management, organized as DATA4LLM and LLM4DATA with a new 'IaaS' data-quality framework.