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

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

Neuron 3321 in Layer 9

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.5055. Min Act: -0.0000

Data Index: 626661 (C4 (Web Text))

Max Activating Token Index: 758

Text #1

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.5102. Min Act: -0.0000

Data Index: 699950 (C4 (Web Text))

Max Activating Token Index: 565

Text #2

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.4253. Min Act: -0.0000

Data Index: 48518 (C4 (Web Text))

Max Activating Token Index: 661

Text #3

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.4437. Min Act: -0.0000

Data Index: 1089363 (C4 (Web Text))

Max Activating Token Index: 844

Text #4

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.4459. Min Act: -0.0000

Data Index: 133367 (C4 (Web Text))

Max Activating Token Index: 962

Text #5

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.4291. Min Act: -0.0000

Data Index: 544991 (C4 (Web Text))

Max Activating Token Index: 426

Text #6

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.4153. Min Act: -0.0000

Data Index: 200323 (C4 (Web Text))

Max Activating Token Index: 846

Text #7

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.4510. Min Act: -0.0000

Data Index: 1302143 (C4 (Web Text))

Max Activating Token Index: 762

Text #8

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.3617. Min Act: -0.0000

Data Index: 1523530 (Python Code)

Max Activating Token Index: 155

Text #9

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.4286. Min Act: -0.0000

Data Index: 1269830 (C4 (Web Text))

Max Activating Token Index: 386

Text #10

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.3672. Min Act: -0.0000

Data Index: 1240167 (C4 (Web Text))

Max Activating Token Index: 210

Text #11

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.3766. Min Act: -0.0000

Data Index: 466580 (C4 (Web Text))

Max Activating Token Index: 834

Text #12

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.3423. Min Act: -0.0000

Data Index: 1128551 (C4 (Web Text))

Max Activating Token Index: 285

Text #13

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.3383. Min Act: -0.0000

Data Index: 565445 (C4 (Web Text))

Max Activating Token Index: 911

Text #14

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.3453. Min Act: -0.0000

Data Index: 1472512 (Python Code)

Max Activating Token Index: 152

Text #15

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.3409. Min Act: -0.0000

Data Index: 1316950 (C4 (Web Text))

Max Activating Token Index: 95

Text #16

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.3649. Min Act: -0.0000

Data Index: 932478 (C4 (Web Text))

Max Activating Token Index: 355

Text #17

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.3110. Min Act: -0.0000

Data Index: 1489415 (Python Code)

Max Activating Token Index: 160

Text #18

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.3169. Min Act: -0.0000

Data Index: 1196733 (C4 (Web Text))

Max Activating Token Index: 795

Text #19

Max Range: 0.5102. Min Range: -0.5102

Max Act: 0.3398. Min Act: -0.0000

Data Index: 697179 (C4 (Web Text))

Max Activating Token Index: 729