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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Recently, various studies have leveraged Large Language Models (LLMs) to helpdecision-making and planning in environments, and try to align the LLMs'knowledge with the world conditions. Nonetheless, the capacity of LLMs tocontinuously acquire environmental knowledge and adapt in an open world remainsuncertain. In this paper, we propose an approach to spur LLMs to explore theopen world, gather experiences, and learn to improve their task-solvingcapabilities. In this approach, a multi-round feedback-revision mechanism isutilized to encourage LLMs to actively select appropriate revision actionsguided by feedback information from the environment. This facilitatesexploration and enhances the model's performance. Besides, we integratesub-task relabeling to assist LLMs in maintaining consistency in sub-taskplanning and help the model learn the combinatorial nature between tasks,enabling it to complete a wider range of tasks through training based on theacquired exploration experiences. By evaluation in Minecraft, an open-endedsandbox world, we demonstrate that our approach LLaMA-Rider enhances theefficiency of the LLM in exploring the environment, and effectively improvesthe LLM's ability to accomplish more tasks through fine-tuning with merely 1.3kinstances of collected data, showing minimal training costs compared to thebaseline using reinforcement learning.

LLaMA Rider: Spurring Large Language Models to Explore the Open World