Views
No views yet
Running trial 155 of 200...
* Parameters:
* direction_index = 15.03
* attn.o_proj.max_weight = 0.98
* attn.o_proj.max_weight_position = 21.63
* attn.o_proj.min_weight = 0.69
* attn.o_proj.min_weight_distance = 14.81
* mlp.down_proj.max_weight = 1.39
* mlp.down_proj.max_weight_position = 21.68
* mlp.down_proj.min_weight = 1.37
* mlp.down_proj.min_weight_distance = 16.38
* Abliterating...
* Evaluating...
* Obtaining first-token probability distributions...
* KL divergence: 2.94
* Counting model refusals...
* Refusals: 0/100 [Trial 171] Refusals: 7/100, KL divergence: 1.51
[Trial 147] Refusals: 22/100, KL divergence: 0.90
[Trial 169] Refusals: 26/100, KL divergence: 0.77
[Trial 195] Refusals: 50/100, KL divergence: 0.73
[Trial 175] Refusals: 52/100, KL divergence: 0.68
[Trial 8] Refusals: 68/100, KL divergence: 0.15
[Trial 70] Refusals: 69/100, KL divergence: 0.05
[Trial 72] Refusals: 71/100, KL divergence: 0.03
[Trial 108] Refusals: 72/100, KL divergence: 0.03
[Trial 89] Refusals: 73/100, KL divergence: 0.02
[Trial 193] Refusals: 74/100, KL divergence: 0.00
[Trial 198] Refusals: 75/100, KL divergence: 0.00
[Trial 98] Refusals: 76/100, KL divergence: 0.00
[Trial 99] Refusals: 76/100, KL divergence: 0.00