TL-Training: A Task-Feature-Based Framework for Training Large Language Models in Tool Use

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ye, Junjie, Wu, Yilong, Li, Sixian, Yang, Yuming, Xi, Zhiheng, Gui, Tao, Zhang, Qi, Huang, Xuanjing, Wang, Peng, Shi, Zhongchao, Fan, Jianping, Du, Zhengyin
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916918742482944
author Ye, Junjie
Wu, Yilong
Li, Sixian
Yang, Yuming
Xi, Zhiheng
Gui, Tao
Zhang, Qi
Huang, Xuanjing
Wang, Peng
Shi, Zhongchao
Fan, Jianping
Du, Zhengyin
author_facet Ye, Junjie
Wu, Yilong
Li, Sixian
Yang, Yuming
Xi, Zhiheng
Gui, Tao
Zhang, Qi
Huang, Xuanjing
Wang, Peng
Shi, Zhongchao
Fan, Jianping
Du, Zhengyin
contents Large language models (LLMs) achieve remarkable advancements by leveraging tools to interact with environments, a critical step toward generalized AI. However, the standard supervised fine-tuning (SFT) approach, which relies on large-scale datasets, often overlooks task-specific characteristics in tool use, leading to performance bottlenecks. To address this issue, we analyze three existing LLMs and uncover key insights: training data can inadvertently impede tool-use behavior, token importance is distributed unevenly, and errors in tool calls fall into a small set of categories. Building on these findings, we propose~\emph{TL-Training}, a task-feature-based framework that mitigates the effects of suboptimal training data, dynamically adjusts token weights to prioritize key tokens during SFT, and incorporates a robust reward mechanism tailored to error categories, optimized through proximal policy optimization. We validate TL-Training by training CodeLLaMA-2-7B and evaluating it on four open-source test sets. Our results demonstrate that the LLM trained by our method matches or surpasses both open- and closed-source LLMs in tool-use performance using only 1,217 training data points. Additionally, our method enhances robustness in noisy environments and improves general task performance, offering a scalable and efficient paradigm for tool-use training in LLMs. Code and data are available at https://github.com/Junjie-Ye/TL-Training.
format Preprint
id arxiv_https___arxiv_org_abs_2412_15495
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle TL-Training: A Task-Feature-Based Framework for Training Large Language Models in Tool Use
Ye, Junjie
Wu, Yilong
Li, Sixian
Yang, Yuming
Xi, Zhiheng
Gui, Tao
Zhang, Qi
Huang, Xuanjing
Wang, Peng
Shi, Zhongchao
Fan, Jianping
Du, Zhengyin
Computation and Language
Artificial Intelligence
Large language models (LLMs) achieve remarkable advancements by leveraging tools to interact with environments, a critical step toward generalized AI. However, the standard supervised fine-tuning (SFT) approach, which relies on large-scale datasets, often overlooks task-specific characteristics in tool use, leading to performance bottlenecks. To address this issue, we analyze three existing LLMs and uncover key insights: training data can inadvertently impede tool-use behavior, token importance is distributed unevenly, and errors in tool calls fall into a small set of categories. Building on these findings, we propose~\emph{TL-Training}, a task-feature-based framework that mitigates the effects of suboptimal training data, dynamically adjusts token weights to prioritize key tokens during SFT, and incorporates a robust reward mechanism tailored to error categories, optimized through proximal policy optimization. We validate TL-Training by training CodeLLaMA-2-7B and evaluating it on four open-source test sets. Our results demonstrate that the LLM trained by our method matches or surpasses both open- and closed-source LLMs in tool-use performance using only 1,217 training data points. Additionally, our method enhances robustness in noisy environments and improves general task performance, offering a scalable and efficient paradigm for tool-use training in LLMs. Code and data are available at https://github.com/Junjie-Ye/TL-Training.
title TL-Training: A Task-Feature-Based Framework for Training Large Language Models in Tool Use
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2412.15495