mondk/Ultra-Trio-Unfiltered

Hi, here are some datasets that are good for training. They include one for improving benchmark scores, one for better tool usage, and one... well, I’m not quite sure yet. Format: {"messages":[{"role":"user","content":"..."},{"role":"assistant","content":"<think>\n...\n</think>\n\n..."}]} You can follow my progress as I train a MiniCPM-5 2B model on this dataset here: Training Process Ty!

<iframe src="https://silytics.si/embed/item/0ccbd725-3bef-48b9-8dfc-daf945bca04d" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 27 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.