Help train Odysseus.
Odysseus LLM is in the making. Contribute by submitting data for training!
I want to train a model that natively understands Odysseus's tools and knows how to respond inside Odysseus. This will help users with less capable hardware run Odysseus locally, and help anyone use Odysseus more efficiently. But it needs more varied data. You can help by picking a category, creating conversations with a competent model, then submitting them here.
Data gatheredlive
11%dataset mass
12tool regions
activesignal intake
Create training seeds
Pick one tool group.
Focus on one area at a time.
For needed coverage, check here.
Use Odysseus normally.
Create a short chat in agent mode that actually hits the tool group. Two to four turns is fine.
Example: give suggestions to this document -> change paragraph 2 -> make it shorter.
Review the result.
Does it:
- use the right Odysseus tools?
- complete the task cleanly?
- include roughly 3-10 useful tool calls?
- avoid private/real data?
Grab the chat log.
Submit.
Submitting gives 1 point. Accepted samples give 5 points. For large batches, ask for private repo access.