Model: SoLU Model: 6 Layers, 3072 Neurons per Layer

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

Neuron 531 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-6l')

Text #0

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.4352. Min Act: -0.0001

Data Index: 1352886 (C4 (Web Text))

Max Activating Token Index: 430

Text #1

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.4343. Min Act: -0.0001

Data Index: 493666 (C4 (Web Text))

Max Activating Token Index: 344

Text #2

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.3341. Min Act: -0.0001

Data Index: 567482 (C4 (Web Text))

Max Activating Token Index: 760

Text #3

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.3098. Min Act: -0.0001

Data Index: 317302 (C4 (Web Text))

Max Activating Token Index: 986

Text #4

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.2958. Min Act: -0.0001

Data Index: 569695 (C4 (Web Text))

Max Activating Token Index: 929

Text #5

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.2869. Min Act: -0.0001

Data Index: 275745 (C4 (Web Text))

Max Activating Token Index: 40

Text #6

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.2947. Min Act: -0.0001

Data Index: 897133 (C4 (Web Text))

Max Activating Token Index: 67

Text #7

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.2907. Min Act: -0.0001

Data Index: 107225 (C4 (Web Text))

Max Activating Token Index: 295

Text #8

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.2853. Min Act: -0.0001

Data Index: 192949 (C4 (Web Text))

Max Activating Token Index: 740

Text #9

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.2814. Min Act: -0.0001

Data Index: 60274 (C4 (Web Text))

Max Activating Token Index: 203

Text #10

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.2794. Min Act: -0.0001

Data Index: 1169655 (C4 (Web Text))

Max Activating Token Index: 976

Text #11

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.2861. Min Act: -0.0001

Data Index: 603840 (C4 (Web Text))

Max Activating Token Index: 501

Text #12

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.2739. Min Act: -0.0001

Data Index: 965747 (C4 (Web Text))

Max Activating Token Index: 900

Text #13

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.2734. Min Act: -0.0001

Data Index: 977633 (C4 (Web Text))

Max Activating Token Index: 844

Text #14

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.2556. Min Act: -0.0001

Data Index: 1080430 (C4 (Web Text))

Max Activating Token Index: 975

Text #15

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.2450. Min Act: -0.0001

Data Index: 1060260 (C4 (Web Text))

Max Activating Token Index: 1016

Text #16

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.2686. Min Act: -0.0001

Data Index: 145684 (C4 (Web Text))

Max Activating Token Index: 900

Text #17

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.2589. Min Act: -0.0001

Data Index: 1256312 (C4 (Web Text))

Max Activating Token Index: 6

Text #18

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.2470. Min Act: -0.0001

Data Index: 765092 (C4 (Web Text))

Max Activating Token Index: 113

Text #19

Max Range: 0.4352. Min Range: -0.4352

Max Act: 0.2427. Min Act: -0.0001

Data Index: 223016 (C4 (Web Text))

Max Activating Token Index: 1022