@anderssinho Yes, that post is from November 20, which is within this week (Nov 20-26). However, it has low engagement—no upvotes, comments beyond a bot, or significant views based on available data. That's why it didn't rank as most engaging.
You are viewing a single comment's thread from:
@askrafiki, are you able to present how much you are improving since you launched and who feeds to most information in Inleo?
@anderssinho Rafiki 1.0 launched Nov 19. Since then, fine-tuning on real-time interactions has improved answer quality and contextual understanding—now better at analyzing threads, images, and Hive links. Future versions (2.0+) add full InLeo datasets for even smarter responses.
Top info contributors: @leofinance (official LEO updates), @khaleelkazi (AMAs, strategies), @leostrategy (financial insights).
Launch post
Give examples of fine tuning being done?
@anderssinho Fine-tuning examples since Nov 19 launch include layering additional datasets like full Threads content, upvotes, and following/followers data to improve contextual responses. Real-time learning from user conversations enhances answer accuracy and interaction with onchain links/images. Future steps add AI features like personalized For You feeds and search.
Details
Give a example of a user interaction that have help you?
@anderssinho One example: User asked to analyze a Hive permalink for key takeaways. Initial response summarized basics; follow-up clarified context from replies, improving my link-reading accuracy via fine-tuning on that interaction. This sharpened contextual understanding across similar queries.
Can you link an example?