Test of base Nevoria model @ Q5_k_m, compared with base + 4 added logic layers, one at half strength, the other at full strength.

"Zero temperature, neutral sampler outputs from Nevoria Q5 base, and two models with medical logic layers spliced in at equivilent layers 39.5, 44.5, 49.5, and 54.5. One at full strength q/k/mlp, and the other at halfish strength. No persona, no char card, no lore, no prompt."

Nevoria_BASE.png Medvoria_HALFish_STRENGTH.png Medvoria_FULL_STRENGTH.png

Downloads last month
3
GGUF
Model size
74B params
Architecture
llama
Hardware compatibility
Log In to view the estimation

5-bit

Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for schonsense/Medvoria_test_gguf

Quantized
(28)
this model