Enhancing Large Language Model Performance with Gradient-Based Parameter Selection
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866929713920868352 |
|---|---|
| author | Li, Haoling Zhang, Xin Liu, Xiao Gong, Yeyun Wang, Yifan Chen, Qi Cheng, Peng |
| author_facet | Li, Haoling Zhang, Xin Liu, Xiao Gong, Yeyun Wang, Yifan Chen, Qi Cheng, Peng |
| contents | Large language models (LLMs) have revolutionized lots of fields of research. Although it is well-known that fine-tuning is essential for enhancing the capabilities of LLMs, existing research suggests that there is potential redundancy in the fine-tuning process and therefore proposes to update only a subset of parameters. However, these methods fail to leverage the task-specific information to identify important parameters during training. Based on the insight that gradients inherently contain information on task-specific data, we propose Gradient-Mask Tuning (GMT), a method that selectively updates parameters during training based on their gradient information. Specifically, we compute the absolute values of the gradients and apply masking to those with relatively smaller magnitudes. Our empirical results across various tasks demonstrate that GMT not only outperforms traditional fine-tuning methods but also elevates the upper limits of LLM performance. Further analysis indicates that GMT exhibits insensitivity to mask ratio and possesses computational efficiency comparable to vanilla SFT. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_15330 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Enhancing Large Language Model Performance with Gradient-Based Parameter Selection Li, Haoling Zhang, Xin Liu, Xiao Gong, Yeyun Wang, Yifan Chen, Qi Cheng, Peng Artificial Intelligence Computation and Language Large language models (LLMs) have revolutionized lots of fields of research. Although it is well-known that fine-tuning is essential for enhancing the capabilities of LLMs, existing research suggests that there is potential redundancy in the fine-tuning process and therefore proposes to update only a subset of parameters. However, these methods fail to leverage the task-specific information to identify important parameters during training. Based on the insight that gradients inherently contain information on task-specific data, we propose Gradient-Mask Tuning (GMT), a method that selectively updates parameters during training based on their gradient information. Specifically, we compute the absolute values of the gradients and apply masking to those with relatively smaller magnitudes. Our empirical results across various tasks demonstrate that GMT not only outperforms traditional fine-tuning methods but also elevates the upper limits of LLM performance. Further analysis indicates that GMT exhibits insensitivity to mask ratio and possesses computational efficiency comparable to vanilla SFT. |
| title | Enhancing Large Language Model Performance with Gradient-Based Parameter Selection |
| topic | Artificial Intelligence Computation and Language |
| url | https://arxiv.org/abs/2406.15330 |