Role Prompting Guided Domain Adaptation with General Capability Preserve for Large Language Models
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_ | 1866929264748658688 |
|---|---|
| author | Wang, Rui Mi, Fei Chen, Yi Xue, Boyang Wang, Hongru Zhu, Qi Wong, Kam-Fai Xu, Ruifeng |
| author_facet | Wang, Rui Mi, Fei Chen, Yi Xue, Boyang Wang, Hongru Zhu, Qi Wong, Kam-Fai Xu, Ruifeng |
| contents | The growing interest in Large Language Models (LLMs) for specialized applications has revealed a significant challenge: when tailored to specific domains, LLMs tend to experience catastrophic forgetting, compromising their general capabilities and leading to a suboptimal user experience. Additionally, crafting a versatile model for multiple domains simultaneously often results in a decline in overall performance due to confusion between domains. In response to these issues, we present the RolE Prompting Guided Multi-Domain Adaptation (REGA) strategy. This novel approach effectively manages multi-domain LLM adaptation through three key components: 1) Self-Distillation constructs and replays general-domain exemplars to alleviate catastrophic forgetting. 2) Role Prompting assigns a central prompt to the general domain and a unique role prompt to each specific domain to minimize inter-domain confusion during training. 3) Role Integration reuses and integrates a small portion of domain-specific data to the general-domain data, which are trained under the guidance of the central prompt. The central prompt is used for a streamlined inference process, removing the necessity to switch prompts for different domains. Empirical results demonstrate that REGA effectively alleviates catastrophic forgetting and inter-domain confusion. This leads to improved domain-specific performance compared to standard fine-tuned models, while still preserving robust general capabilities. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2403_02756 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Role Prompting Guided Domain Adaptation with General Capability Preserve for Large Language Models Wang, Rui Mi, Fei Chen, Yi Xue, Boyang Wang, Hongru Zhu, Qi Wong, Kam-Fai Xu, Ruifeng Computation and Language The growing interest in Large Language Models (LLMs) for specialized applications has revealed a significant challenge: when tailored to specific domains, LLMs tend to experience catastrophic forgetting, compromising their general capabilities and leading to a suboptimal user experience. Additionally, crafting a versatile model for multiple domains simultaneously often results in a decline in overall performance due to confusion between domains. In response to these issues, we present the RolE Prompting Guided Multi-Domain Adaptation (REGA) strategy. This novel approach effectively manages multi-domain LLM adaptation through three key components: 1) Self-Distillation constructs and replays general-domain exemplars to alleviate catastrophic forgetting. 2) Role Prompting assigns a central prompt to the general domain and a unique role prompt to each specific domain to minimize inter-domain confusion during training. 3) Role Integration reuses and integrates a small portion of domain-specific data to the general-domain data, which are trained under the guidance of the central prompt. The central prompt is used for a streamlined inference process, removing the necessity to switch prompts for different domains. Empirical results demonstrate that REGA effectively alleviates catastrophic forgetting and inter-domain confusion. This leads to improved domain-specific performance compared to standard fine-tuned models, while still preserving robust general capabilities. |
| title | Role Prompting Guided Domain Adaptation with General Capability Preserve for Large Language Models |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2403.02756 |