Sat Grewal PRO
satgeze
AI & ML interests
None yet
Recent Activity
posted an update about 11 hours ago
Two abliterated Ornith-1.0 builds, now public. Refusal removed with a single rank-1 weight edit on the MLP down-projections, direction extracted from matched harmful/harmless pairs. Measured, not claimed: 35B at **0% refusal, KL 0.036, no capability loss** (GSM8K 96.0 vs 94.0); 9B at **2% residual, KL 0.070**, an honest 2.5-point tradeoff. Verified on a real-world gate (no language drift, output integrity, real engagement), not just a refusal count. GGUF, Q8_0/Q4_K_M.
https://huggingface.co/satgeze/Ornith-1.0-35B-abliterated-SatGeZe
https://huggingface.co/satgeze/Ornith-1.0-9B-abliterated-SatGeZe
updated a model about 12 hours ago
satgeze/Ornith-1.0-35B-abliterated-SatGeZe published a model about 13 hours ago
satgeze/Ornith-1.0-35B-abliterated-SatGeZeOrganizations
None yet