Model: GELU Model: 4 Layers, 2048 Neurons per Layer

Dataset: 80% C4 (Web Text) and 20% Python Code

Neuron 18 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-4l')

Text #0

Max Range: 5.3162. Min Range: -5.3162

Max Act: 5.3162. Min Act: -0.1700

Data Index: 156915 (C4 (Web Text))

Max Activating Token Index: 132

Text #1

Max Range: 5.3162. Min Range: -5.3162

Max Act: 5.2850. Min Act: -0.1700

Data Index: 813366 (C4 (Web Text))

Max Activating Token Index: 215

Text #2

Max Range: 5.3162. Min Range: -5.3162

Max Act: 5.2225. Min Act: -0.1700

Data Index: 805605 (C4 (Web Text))

Max Activating Token Index: 378

Text #3

Max Range: 5.3162. Min Range: -5.3162

Max Act: 5.2537. Min Act: -0.1700

Data Index: 966884 (C4 (Web Text))

Max Activating Token Index: 903

Text #4

Max Range: 5.3162. Min Range: -5.3162

Max Act: 5.1600. Min Act: -0.1700

Data Index: 664181 (C4 (Web Text))

Max Activating Token Index: 566

Text #5

Max Range: 5.3162. Min Range: -5.3162

Max Act: 5.1600. Min Act: -0.1700

Data Index: 1192817 (C4 (Web Text))

Max Activating Token Index: 400

Text #6

Max Range: 5.3162. Min Range: -5.3162

Max Act: 5.1287. Min Act: -0.1700

Data Index: 178500 (C4 (Web Text))

Max Activating Token Index: 261

Text #7

Max Range: 5.3162. Min Range: -5.3162

Max Act: 5.0975. Min Act: -0.1700

Data Index: 1107584 (C4 (Web Text))

Max Activating Token Index: 957

Text #8

Max Range: 5.3162. Min Range: -5.3162

Max Act: 5.0350. Min Act: -0.1700

Data Index: 682031 (C4 (Web Text))

Max Activating Token Index: 157

Text #9

Max Range: 5.3162. Min Range: -5.3162

Max Act: 5.0350. Min Act: -0.1700

Data Index: 480961 (C4 (Web Text))

Max Activating Token Index: 76

Text #10

Max Range: 5.3162. Min Range: -5.3162

Max Act: 5.0350. Min Act: -0.1700

Data Index: 977469 (C4 (Web Text))

Max Activating Token Index: 424

Text #11

Max Range: 5.3162. Min Range: -5.3162

Max Act: 5.0662. Min Act: -0.1700

Data Index: 494832 (C4 (Web Text))

Max Activating Token Index: 202

Text #12

Max Range: 5.3162. Min Range: -5.3162

Max Act: 5.0350. Min Act: -0.1700

Data Index: 1147624 (C4 (Web Text))

Max Activating Token Index: 868

Text #13

Max Range: 5.3162. Min Range: -5.3162

Max Act: 5.0037. Min Act: -0.1699

Data Index: 319690 (C4 (Web Text))

Max Activating Token Index: 244

Text #14

Max Range: 5.3162. Min Range: -5.3162

Max Act: 4.9725. Min Act: -0.1700

Data Index: 300401 (C4 (Web Text))

Max Activating Token Index: 94

Text #15

Max Range: 5.3162. Min Range: -5.3162

Max Act: 4.9412. Min Act: -0.1700

Data Index: 117983 (C4 (Web Text))

Max Activating Token Index: 86

Text #16

Max Range: 5.3162. Min Range: -5.3162

Max Act: 4.8787. Min Act: -0.1700

Data Index: 801167 (C4 (Web Text))

Max Activating Token Index: 933

Text #17

Max Range: 5.3162. Min Range: -5.3162

Max Act: 4.9412. Min Act: -0.1700

Data Index: 1125829 (C4 (Web Text))

Max Activating Token Index: 921

Text #18

Max Range: 5.3162. Min Range: -5.3162

Max Act: 4.9100. Min Act: -0.1700

Data Index: 100332 (C4 (Web Text))

Max Activating Token Index: 405

Text #19

Max Range: 5.3162. Min Range: -5.3162

Max Act: 4.9725. Min Act: -0.1700

Data Index: 111936 (C4 (Web Text))

Max Activating Token Index: 914