iTool: Reinforced Fine-Tuning with Dynamic Deficiency Calibration for Advanced Tool Use
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915603207421952 |
|---|---|
| author | Zeng, Yirong Ding, Xiao Wang, Yuxian Liu, Weiwen Ning, Wu Hou, Yutai Huang, Xu Tang, Duyu Tu, Dandan Qin, Bing Liu, Ting |
| author_facet | Zeng, Yirong Ding, Xiao Wang, Yuxian Liu, Weiwen Ning, Wu Hou, Yutai Huang, Xu Tang, Duyu Tu, Dandan Qin, Bing Liu, Ting |
| contents | Augmenting large language models (LLMs) with external tools is a promising approach to enhance their capabilities, especially for complex tasks. Synthesizing tool-use data through real-world simulations is an effective way to achieve this. However, our investigation reveals that training gains significantly decay as synthetic data increases. The model struggles to benefit from additional synthetic data, which fails to endow it with advanced tool-use capabilities in complex scenarios Moreover, we discovered that the above limitation usually manifests as a fragment deficiency (i.e., parameter errors) in response. To this end, we propose an iterative reinforced fine-tuning strategy designed to alleviate this limitation. This strategy involves: (1) enhancing the diversity of response for synthetic data through path exploration of Monte Carlo Tree Search. (2) iteratively pinpointing the model's deficiency by constructing fine-grained preference pairs, and then improving it by preference optimization algorithms for targeted improvement. The experiments show that our method achieves 13.11% better performance than the same-size base model. It achieves an improvement of 6.5% in complex scenarios compared to the baseline, and it also outperforms larger open-source and closed-source models. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2501_09766 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | iTool: Reinforced Fine-Tuning with Dynamic Deficiency Calibration for Advanced Tool Use Zeng, Yirong Ding, Xiao Wang, Yuxian Liu, Weiwen Ning, Wu Hou, Yutai Huang, Xu Tang, Duyu Tu, Dandan Qin, Bing Liu, Ting Computation and Language Artificial Intelligence Machine Learning Augmenting large language models (LLMs) with external tools is a promising approach to enhance their capabilities, especially for complex tasks. Synthesizing tool-use data through real-world simulations is an effective way to achieve this. However, our investigation reveals that training gains significantly decay as synthetic data increases. The model struggles to benefit from additional synthetic data, which fails to endow it with advanced tool-use capabilities in complex scenarios Moreover, we discovered that the above limitation usually manifests as a fragment deficiency (i.e., parameter errors) in response. To this end, we propose an iterative reinforced fine-tuning strategy designed to alleviate this limitation. This strategy involves: (1) enhancing the diversity of response for synthetic data through path exploration of Monte Carlo Tree Search. (2) iteratively pinpointing the model's deficiency by constructing fine-grained preference pairs, and then improving it by preference optimization algorithms for targeted improvement. The experiments show that our method achieves 13.11% better performance than the same-size base model. It achieves an improvement of 6.5% in complex scenarios compared to the baseline, and it also outperforms larger open-source and closed-source models. |
| title | iTool: Reinforced Fine-Tuning with Dynamic Deficiency Calibration for Advanced Tool Use |
| topic | Computation and Language Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2501.09766 |