Official: 'TianGong GPT' Released!

CN: https://gpt.tiangong.world/
EN: https://gpt.tiangong.world/en

1. About Tiangong GPT

1.1 Tiangong GPT

• It is an intelligent analysis tool for academic papers in the field of environment and sustainable development.
• Developed by Professor Ming Xu's Environmental Data Science and Systems Engineering team at Tsinghua University's School of Environment.
• The tool features a chat-based interface in multiple languages to help improve research and work efficiency.
• It is currently in the Research Preview stage, and can be accessed free of charge without the need for login credentials.

1.2 How to Use the Tool

• Step 1: Select and upload a PDF file.
• Step 2: Wait for the file to be analyzed.
• Step 3: Start chatting!

1.3 Some additional information

• TianGong GPT is specifically trained for scientific research in the field of environment and sustainable development. Please select papers within this field.
• Chatting is based on the selected paper, so conversations should be conducted paper by paper.
• For further information, please refer to the beta test instructions provided on the interface.

2. Prompts: How to Chat with TianGong GPT

• TianGong GPT's prompts support various languages.

• Prompts can also be a mix of multiple languages, and GPT will respond based on the primary language of the question.

• Here are some prompts that you can try out:

  • 这篇论文的主要内容是什么
  • 主要创新点有哪些
  • 详细介绍下研究方法
  • 主要的贡献和implications有哪些
  • 论文的Discussion有哪些主要观点？
  • Summarize the introduction of this paper.
  • Provide a literature review of this paper.
  • Present the results of the study.
  • Conclusions from the paper
  • Limitations of this paper
  • Future works suggested in this paper

3. Follow-up

3.1 Model Training and Tool Updates

• Efforts will be continuously made to organize data labeling and optimize the GPT model.
• The model and tool will be updated regularly without any noticeable interruptions, ensuring an improved user experience.

3.2 Testing and Feedback

• TianGong GPT is still in the Research Preview beta testing stage and welcomes feedback on any issues.

• In situations where there is a high volume of users and server load, access issues may occur. In such cases, please refresh or reopen the page.

• If a PDF file cannot be read or other abnormalities occur, please:

  • Confirm that the PDF file is a text content version, not a scanned image.
  • Refresh the page, close and reopen the browser, or clear the browser cache.

• Any issues, screenshots, and problematic PDF files can be reported to [email protected].

4. Exploring More Possibilities

• We support local deployment of TianGong GPT to enhance user experience.
• We welcome collaboration to explore more possible application scenarios for TianGong GPT.