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

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

Neuron 333 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0444. Min Act: -0.0001

Data Index: 122230 (C4 (Web Text))

Max Activating Token Index: 94

Text #1

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0452. Min Act: -0.0002

Data Index: 1021271 (C4 (Web Text))

Max Activating Token Index: 690

Text #2

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0423. Min Act: -0.0002

Data Index: 481736 (C4 (Web Text))

Max Activating Token Index: 738

Text #3

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0418. Min Act: -0.0002

Data Index: 659595 (C4 (Web Text))

Max Activating Token Index: 360

Text #4

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0403. Min Act: -0.0001

Data Index: 1625991 (Python Code)

Max Activating Token Index: 177

Text #5

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0408. Min Act: -0.0001

Data Index: 1361955 (Python Code)

Max Activating Token Index: 251

Text #6

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0428. Min Act: 0.0000

Data Index: 516917 (C4 (Web Text))

Max Activating Token Index: 437

Text #7

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0393. Min Act: -0.0001

Data Index: 104023 (C4 (Web Text))

Max Activating Token Index: 816

Text #8

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0391. Min Act: -0.0002

Data Index: 103246 (C4 (Web Text))

Max Activating Token Index: 879

Text #9

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0375. Min Act: -0.0001

Data Index: 627353 (C4 (Web Text))

Max Activating Token Index: 61

Text #10

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0385. Min Act: -0.0001

Data Index: 841951 (C4 (Web Text))

Max Activating Token Index: 554

Text #11

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0389. Min Act: -0.0002

Data Index: 170773 (C4 (Web Text))

Max Activating Token Index: 429

Text #12

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0377. Min Act: -0.0002

Data Index: 486972 (C4 (Web Text))

Max Activating Token Index: 20

Text #13

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0359. Min Act: -0.0001

Data Index: 729505 (C4 (Web Text))

Max Activating Token Index: 135

Text #14

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0349. Min Act: -0.0001

Data Index: 92600 (C4 (Web Text))

Max Activating Token Index: 8

Text #15

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0362. Min Act: -0.0001

Data Index: 1559297 (Python Code)

Max Activating Token Index: 989

Text #16

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0359. Min Act: -0.0001

Data Index: 1347064 (C4 (Web Text))

Max Activating Token Index: 122

Text #17

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0366. Min Act: -0.0002

Data Index: 1042618 (C4 (Web Text))

Max Activating Token Index: 755

Text #18

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0353. Min Act: -0.0001

Data Index: 846199 (C4 (Web Text))

Max Activating Token Index: 144

Text #19

Max Range: 0.0452. Min Range: -0.0452

Max Act: 0.0345. Min Act: -0.0001

Data Index: 1457254 (Python Code)

Max Activating Token Index: 51