Model: SoLU Model: 10 Layers, 5120 Neurons per Layer

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

Neuron 3552 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.4790. Min Act: 0.0000

Data Index: 402467 (C4 (Web Text))

Max Activating Token Index: 696

Text #1

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.3695. Min Act: -0.0000

Data Index: 1522046 (Python Code)

Max Activating Token Index: 618

Text #2

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.3293. Min Act: -0.0000

Data Index: 1366116 (Python Code)

Max Activating Token Index: 235

Text #3

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.4247. Min Act: -0.0000

Data Index: 292195 (C4 (Web Text))

Max Activating Token Index: 842

Text #4

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.3877. Min Act: -0.0000

Data Index: 1638514 (Python Code)

Max Activating Token Index: 982

Text #5

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.3615. Min Act: -0.0000

Data Index: 807684 (C4 (Web Text))

Max Activating Token Index: 603

Text #6

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.3759. Min Act: -0.0000

Data Index: 95184 (C4 (Web Text))

Max Activating Token Index: 637

Text #7

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.3825. Min Act: -0.0000

Data Index: 1417596 (Python Code)

Max Activating Token Index: 634

Text #8

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.2959. Min Act: -0.0000

Data Index: 926995 (C4 (Web Text))

Max Activating Token Index: 353

Text #9

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.3491. Min Act: -0.0000

Data Index: 1398895 (Python Code)

Max Activating Token Index: 874

Text #10

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.3112. Min Act: 0.0000

Data Index: 1325248 (C4 (Web Text))

Max Activating Token Index: 66

Text #11

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.3521. Min Act: -0.0000

Data Index: 1539789 (Python Code)

Max Activating Token Index: 854

Text #12

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.3310. Min Act: -0.0000

Data Index: 1479630 (Python Code)

Max Activating Token Index: 478

Text #13

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.3280. Min Act: -0.0000

Data Index: 25552 (C4 (Web Text))

Max Activating Token Index: 991

Text #14

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.3201. Min Act: -0.0000

Data Index: 1396139 (Python Code)

Max Activating Token Index: 934

Text #15

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.3079. Min Act: -0.0000

Data Index: 519782 (C4 (Web Text))

Max Activating Token Index: 487

Text #16

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.3058. Min Act: -0.0000

Data Index: 1556511 (Python Code)

Max Activating Token Index: 768

Text #17

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.3417. Min Act: -0.0000

Data Index: 198049 (C4 (Web Text))

Max Activating Token Index: 224

Text #18

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.3159. Min Act: -0.0000

Data Index: 1117834 (C4 (Web Text))

Max Activating Token Index: 135

Text #19

Max Range: 0.4790. Min Range: -0.4790

Max Act: 0.3099. Min Act: -0.0000

Data Index: 1558119 (Python Code)

Max Activating Token Index: 841