Spaces:

---

DontPlanToEnd

/

UGI-Leaderboard

like 1.81k

Running

Please evaluate my model:
https://huggingface.co/morikomorizz/GRM-2.6-Plus-Primal

27B parameter abliterated model based on OrionLLM/GRM-2.6-Plus (Qwen3.6-27B architecture).
Uses Heretic-ARA abliteration. KL divergence: 0.0678. Refusals: 6/100.
Multimodal (vision + text).