English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Large Language Models (LLMs) have demonstrated remarkable progress inutilizing tools, but their closed-source nature and high inference costs poselimitations on their adaptability, necessitating a valid method that leveragessmaller, open-sourced models. In this paper, we introduce Toolink, acomprehensive framework that performs task-solving by first creating a toolkitand then integrating the planning and calling of tools through achain-of-solving (CoS) approach. We first validate the efficacy of Toolink inharnessing the model's creativity and CoS ability on ChatGPT. Subsequently, wecurate CoS-GPT, a chain-of-solving dataset designed for tool-using, andfinetune the LLaMA-7B model. It results in LLaMA-CoS, a powerful open-sourcemodel with advanced tool-planning and tool-calling capabilities. Evaluation ondiverse tasks from BIG-bench demonstrates its CoS ability matches that ofChatGPT while its performance surpasses the chain-of-thought approach. Furtherstudies highlight the generalization of LLaMA-CoS to unseen tasks and showcaseits capability in using toolkits not explicitly tailored for the target task,affirming its robustness in real-world scenarios. All codes and data arereleased.

Toolink: Linking Toolkit Creation and Using through Chain-of-Solving on Open-Source Model