Gain Insights from Recent Published Papers

In this demo I will use Helix to learn from recently published academic papers and provide insight. I will search arxiv, a popular openly accessible paper website, for a recent technical topic. Then I will take those results and ask Helix to fine tune a large language model based upon them. Then I will be able to gain insight from these articles.

Steps

1. Start an inference session on Helix
2. Type a question relating to the topic of interest and note how incorrect it is, e.g.:
   • How are LLMs improving 6G communications networks?
   • What components make up the proposed multi-agent system for 6G communications?
   • What is Multi-agent Data Retrieval in multi-agent systems for 6G communications?
   • How does Multi-agent Data Retrieval work in multi-agent systems for 6G communications?
   • Generate 3 blog post titles based upon recent papers about the use of LLMs in 6G communication networks.
3. Start a finetune session on Helix
   • Search Arxiv for papers of interest, e.g.:
   • https://arxiv.org/search/?query=+6G+communications+llms&searchtype=all
   • Copy the Links of the PDFs and paste them into Helix’s “Add link” section, one by one. E.g.:
     • https://arxiv.org/pdf/2312.07850
     • https://arxiv.org/pdf/2311.05842
     • https://arxiv.org/pdf/2308.09376
   • Hit finetune
4. When finished, re-ask the same questions:
   • How are LLMs improving 6G communications networks?
   • What components make up the proposed multi-agent system for 6G communications?
   • What is Multi-agent Data Retrieval in multi-agent systems for 6G communications?
   • How does Multi-agent Data Retrieval work in multi-agent systems for 6G communications?
   • Generate 3 blog post titles based upon recent papers about the use of LLMs in 6G communication networks.

Example Sessions

• https://app.tryhelix.ai/session/f680a35f-6e99-4188-96d4-9582dbf2ef56