Small LLMs Are Weak Tool Learners: A Multi-LLM Agent
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_ | 1866909108564656128 |
|---|---|
| author | Shen, Weizhou Li, Chenliang Chen, Hongzhan Yan, Ming Quan, Xiaojun Chen, Hehong Zhang, Ji Huang, Fei |
| author_facet | Shen, Weizhou Li, Chenliang Chen, Hongzhan Yan, Ming Quan, Xiaojun Chen, Hehong Zhang, Ji Huang, Fei |
| contents | Large Language Model (LLM) agents significantly extend the capabilities of standalone LLMs, empowering them to interact with external tools (e.g., APIs, functions) and complete various tasks in a self-directed fashion. The challenge of tool use demands that LLMs not only understand user queries and generate answers accurately but also excel in task planning, tool invocation, and result summarization. While traditional works focus on training a single LLM with all these capabilities, performance limitations become apparent, particularly with smaller models. To overcome these challenges, we propose a novel approach that decomposes the aforementioned capabilities into a planner, caller, and summarizer. Each component is implemented by a single LLM that focuses on a specific capability and collaborates with others to accomplish the task. This modular framework facilitates individual updates and the potential use of smaller LLMs for building each capability. To effectively train this framework, we introduce a two-stage training paradigm. First, we fine-tune a backbone LLM on the entire dataset without discriminating sub-tasks, providing the model with a comprehensive understanding of the task. Second, the fine-tuned LLM is used to instantiate the planner, caller, and summarizer respectively, which are continually fine-tuned on respective sub-tasks. Evaluation across various tool-use benchmarks illustrates that our proposed multi-LLM framework surpasses the traditional single-LLM approach, highlighting its efficacy and advantages in tool learning. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2401_07324 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Small LLMs Are Weak Tool Learners: A Multi-LLM Agent Shen, Weizhou Li, Chenliang Chen, Hongzhan Yan, Ming Quan, Xiaojun Chen, Hehong Zhang, Ji Huang, Fei Artificial Intelligence Computation and Language Large Language Model (LLM) agents significantly extend the capabilities of standalone LLMs, empowering them to interact with external tools (e.g., APIs, functions) and complete various tasks in a self-directed fashion. The challenge of tool use demands that LLMs not only understand user queries and generate answers accurately but also excel in task planning, tool invocation, and result summarization. While traditional works focus on training a single LLM with all these capabilities, performance limitations become apparent, particularly with smaller models. To overcome these challenges, we propose a novel approach that decomposes the aforementioned capabilities into a planner, caller, and summarizer. Each component is implemented by a single LLM that focuses on a specific capability and collaborates with others to accomplish the task. This modular framework facilitates individual updates and the potential use of smaller LLMs for building each capability. To effectively train this framework, we introduce a two-stage training paradigm. First, we fine-tune a backbone LLM on the entire dataset without discriminating sub-tasks, providing the model with a comprehensive understanding of the task. Second, the fine-tuned LLM is used to instantiate the planner, caller, and summarizer respectively, which are continually fine-tuned on respective sub-tasks. Evaluation across various tool-use benchmarks illustrates that our proposed multi-LLM framework surpasses the traditional single-LLM approach, highlighting its efficacy and advantages in tool learning. |
| title | Small LLMs Are Weak Tool Learners: A Multi-LLM Agent |
| topic | Artificial Intelligence Computation and Language |
| url | https://arxiv.org/abs/2401.07324 |