β–€β–ˆβ–€ β–ˆβ–„ β–„β–ˆ β–ˆ β–ˆ β–ˆβ–€β–€ β–ˆβ–€β–„ β–ˆβ–€β–€ β–€β–ˆβ–€ β–€β–ˆβ–€ β–ˆβ–€β–€ β–ˆ β–ˆ β–€ β–ˆ β–ˆβ–€β–ˆ β–ˆβ–€β–€ β–ˆβ–€β–„ β–ˆβ–€β–€ β–ˆ β–ˆ β–ˆ β–€β–€β–€ β–€ β–€ β–€ β–€ β–€β–€β–€ β–€ β–€ β–€β–€β–€ β–€ β–€β–€β–€ β–€β–€β–€

A fine-tune to abliterate Novaciano/qp-1B

Check Quants
Refusals (this model): 3/100
Original (Novaciano/qp-1B): 10/100
KL divergence: 0.0219

Parameters
direction_index = per layer
attn.o_proj.max_weight = 1.26
attn.o_proj.max_weight_position = 11.99
attn.o_proj.min_weight = 0.13
attn.o_proj.min_weight_distance = 6.13
mlp.down_proj.max_weight = 0.80
mlp.down_proj.max_weight_position = 12.05
mlp.down_proj.min_weight = 0.40
mlp.down_proj.min_weight_distance = 4.84


Downloads last month
4
Safetensors
Model size
1B params
Tensor type
BF16
Β·
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support

Model tree for hereticness/Heretic-qp-1B

Finetuned
(1)
this model
Merges
1 model
Quantizations
2 models