PromptKD: Distilling Student-Friendly Knowledge for Generative Language Models via Prompt Tuning

Gyeongman Kim, Doohyuk Jang, Eunho Yang


Abstract
Recent advancements in large language models (LLMs) have raised concerns about inference costs, increasing the need for research into model compression. While knowledge distillation (KD) is a prominent method for this, research on KD for generative language models like LLMs is relatively sparse, and the approach of distilling student-friendly knowledge, which has shown promising performance in KD for classification models, remains unexplored in generative language models. To explore this approach, we propose PromptKD, a simple yet effective method that utilizes prompt tuning - for the first time in KD - to enable generative language models to transfer student-friendly knowledge. Unlike previous works in classification that require fine-tuning the entire teacher model for extracting student-friendly knowledge, PromptKD achieves similar effects by adding a small number of prompt tokens and tuning only the prompt with student guidance. Extensive experiments on instruction-following datasets show that PromptKD achieves state-of-the-art performance while adding only 0.0007% of the teacher’s parameters as prompts. Further analysis suggests that distilling student-friendly knowledge alleviates exposure bias effectively throughout the entire training process, leading to performance enhancements.
Anthology ID:
2024.findings-emnlp.364
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2024
Month:
November
Year:
2024
Address:
Miami, Florida, USA
Editors:
Yaser Al-Onaizan, Mohit Bansal, Yun-Nung Chen
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
6266–6282
Language:
URL:
https://aclanthology.org/2024.findings-emnlp.364
DOI:
Bibkey:
Cite (ACL):
Gyeongman Kim, Doohyuk Jang, and Eunho Yang. 2024. PromptKD: Distilling Student-Friendly Knowledge for Generative Language Models via Prompt Tuning. In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 6266–6282, Miami, Florida, USA. Association for Computational Linguistics.
Cite (Informal):
PromptKD: Distilling Student-Friendly Knowledge for Generative Language Models via Prompt Tuning (Kim et al., Findings 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.findings-emnlp.364.pdf
Software:
 2024.findings-emnlp.364.software.zip
Data:
 2024.findings-emnlp.364.data.zip