LlamaIndex Webinar: RAPTOR - Tree-Structured Indexing and Retrieval

Поделиться
HTML-код
  • Опубликовано: 28 сен 2024

Комментарии • 7

  • @ai-cowboy
    @ai-cowboy 6 месяцев назад +6

    Adding "DON'T RELY ON PRIOR KNOWLEDGE" in your summarization prompts helped me reduce hallucinations

  • @microburn
    @microburn 4 месяца назад

    Too bad the audio kept cutting out. Good information

  • @marcusturner5285
    @marcusturner5285 6 месяцев назад

    I getting rate limit when using the lama pack

  • @alicexia834
    @alicexia834 7 дней назад

    Great presentation and amazing Q&A. I have question here regarding with updating financial reports. For example, I have 3,000 U.S. listed companies' financial reports for 10 years. As the Q&A part advised, I put the year such as 2010 in every chunk. Do I need to put stock name in every chunk to reduce the hallucination during summarisation since finanical reports follow similar patterns , too?

  • @Darthus
    @Darthus 4 месяца назад

    Man, someone really needed to mute their mic here, otherwise great presentation!

  • @jma7889
    @jma7889 6 месяцев назад +1

    TL;DR

  • @jaapjaap1
    @jaapjaap1 6 месяцев назад

    Very interesting presentation. Thanks!