🧠nanoMoE 10R — Dense 124M (SFT v8)
Dense 124M base + SFT v8 (ChatML) — loaded from Daxamite/10R_Dense_124m.
A GPT-2-small-class assistant: simple instructions, basic QA, and code completion. Optional built-in math (sympy) and web-search tools.
Small model — expect mistakes; not for production use.