Visualization Directory Index

Root
assets
classification\meta-llama\Llama-3.2-1B
EleutherAI\sae-Llama-3.2-1B-131k\MartinThoma\wili_2018
sae-count
sae-min-max
MartinThoma\wili_2018\neuron-count
correlation\meta-llama\Llama-3.2-1B\EleutherAI\sae-Llama-3.2-1B-131k
detection
fuzz
entropy\meta-llama\Llama-3.2-1B\EleutherAI\sae-Llama-3.2-1B-131k
feature_index\Llama-3.2-1B\sae-Llama-3.2-1B-131k
model.layers.0.mlp
Bulgarian
English
Hindi
Japanese
Korean
Portuguese
Thai
Turkish
Vietnamese
model.layers.1.mlp
Bulgarian
Hindi
Italian
Thai
Turkish
model.layers.2.mlp
Chinese
English
German
Hindi
Italian
Portuguese
Spanish
Thai
Turkish
model.layers.3.mlp
Bulgarian
Hindi
Italian
Portuguese
Spanish
Thai
Turkish
Vietnamese
model.layers.4.mlp
Bulgarian
English
Hindi
Italian
Korean
Portuguese
Russian
Thai
Vietnamese
model.layers.5.mlp
Bulgarian
French
German
Hindi
Portuguese
Spanish
Thai
Turkish
Vietnamese
model.layers.6.mlp
Bulgarian
Chinese
Italian
Korean
Russian
Thai
Turkish
model.layers.7.mlp
Bulgarian
Chinese
Hindi
Italian
Korean
Portuguese
Russian
Thai
Turkish
Vietnamese
model.layers.8.mlp
Bulgarian
English
French
German
Hindi
Italian
Japanese
Korean
Portuguese
Russian
Thai
Turkish
Vietnamese
model.layers.9.mlp
Bulgarian
Chinese
French
German
Hindi
Italian
Japanese
Korean
Portuguese
Russian
Thai
Turkish
Vietnamese
model.layers.10.mlp
Bulgarian
Chinese
German
Hindi
Italian
Korean
Portuguese
Russian
Thai
Turkish
Vietnamese
model.layers.11.mlp
Bulgarian
Chinese
French
German
Hindi
Italian
Japanese
Korean
Portuguese
Russian
Spanish
Thai
Turkish
Vietnamese
model.layers.12.mlp
Bulgarian
Chinese
French
German
Hindi
Italian
Japanese
Korean
Portuguese
Russian
Thai
Turkish
Vietnamese
model.layers.13.mlp
Bulgarian
Chinese
French
Hindi
Italian
Japanese
Korean
Portuguese
Russian
Spanish
Thai
Turkish
Vietnamese
model.layers.14.mlp
Bulgarian
Chinese
French
German
Hindi
Italian
Japanese
Korean
Portuguese
Russian
Thai
Turkish
Vietnamese
model.layers.15.mlp
Bulgarian
Chinese
English
German
Hindi
Japanese
Korean
Portuguese
Russian
Thai
Turkish
Vietnamese
lape\meta-llama\Llama-3.2-1B\EleutherAI\sae-Llama-3.2-1B-131k
lape_neuron
sae_features
lape_all
lape_top_1_per_layer_by_entropy
lape_top_1_per_layer_by_freq
lape_top_10_by_entropy
lape_top_10_by_freq
ppl\meta-llama\Llama-3.2-1B\openlanguagedata\flores_plus
neuron_intervention
baseline
min_0.1
min_0.2
min_0.3
min_0.4
plus_0.1
plus_0.2
plus_0.3
plus_0.4
fixed_0
sae_intervention
all\entropy\max
mult_0.1
mult_0.2
mult_0.3
mult_0.4
mult_-0.1
mult_-0.2
mult_-0.3
mult_-0.4
top_1_per_layer
entropy
avg
mult_1
mult_-1
max
mult_0.2
mult_-0.2
freq\avg\mult_-1
top_10\entropy\max
mult_0.2
mult_-0.2
shared_features\meta-llama\Llama-3.2-1B\EleutherAI\sae-Llama-3.2-1B-131k
similarity\meta-llama\Llama-3.2-1B\EleutherAI\sae-Llama-3.2-1B-131k
iou
pearson
umap\meta-llama\Llama-3.2-1B\EleutherAI\sae-Llama-3.2-1B-131k
model.layers.0.mlp
model.layers.1.mlp
model.layers.2.mlp
model.layers.3.mlp
model.layers.4.mlp
model.layers.5.mlp
model.layers.6.mlp
model.layers.7.mlp
model.layers.8.mlp
model.layers.9.mlp
model.layers.10.mlp
model.layers.11.mlp
model.layers.12.mlp
model.layers.13.mlp
model.layers.14.mlp
model.layers.15.mlp