Model: SoLU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 1590 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-3l')

Text #0

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0437. Min Act: -0.0003

Data Index: 1018828 (C4 (Web Text))

Max Activating Token Index: 754

Text #1

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0439. Min Act: -0.0003

Data Index: 10113 (C4 (Web Text))

Max Activating Token Index: 188

Text #2

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0419. Min Act: -0.0003

Data Index: 1344290 (C4 (Web Text))

Max Activating Token Index: 269

Text #3

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0393. Min Act: -0.0003

Data Index: 280259 (C4 (Web Text))

Max Activating Token Index: 878

Text #4

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0412. Min Act: -0.0003

Data Index: 717580 (C4 (Web Text))

Max Activating Token Index: 243

Text #5

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0402. Min Act: -0.0003

Data Index: 490516 (C4 (Web Text))

Max Activating Token Index: 129

Text #6

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0400. Min Act: -0.0003

Data Index: 957822 (C4 (Web Text))

Max Activating Token Index: 780

Text #7

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0390. Min Act: -0.0003

Data Index: 1211049 (C4 (Web Text))

Max Activating Token Index: 321

Text #8

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0385. Min Act: -0.0003

Data Index: 1090046 (C4 (Web Text))

Max Activating Token Index: 124

Text #9

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0379. Min Act: -0.0003

Data Index: 1556488 (Python Code)

Max Activating Token Index: 335

Text #10

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0370. Min Act: -0.0003

Data Index: 283107 (C4 (Web Text))

Max Activating Token Index: 710

Text #11

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0370. Min Act: -0.0003

Data Index: 678755 (C4 (Web Text))

Max Activating Token Index: 221

Text #12

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0377. Min Act: -0.0003

Data Index: 368916 (C4 (Web Text))

Max Activating Token Index: 397

Text #13

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0371. Min Act: -0.0003

Data Index: 287947 (C4 (Web Text))

Max Activating Token Index: 201

Text #14

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0376. Min Act: -0.0003

Data Index: 1433178 (Python Code)

Max Activating Token Index: 262

Text #15

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0360. Min Act: -0.0003

Data Index: 326865 (C4 (Web Text))

Max Activating Token Index: 136

Text #16

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0373. Min Act: -0.0002

Data Index: 12561 (C4 (Web Text))

Max Activating Token Index: 136

Text #17

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0367. Min Act: -0.0003

Data Index: 326130 (C4 (Web Text))

Max Activating Token Index: 130

Text #18

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0367. Min Act: -0.0003

Data Index: 998762 (C4 (Web Text))

Max Activating Token Index: 147

Text #19

Max Range: 0.0439. Min Range: -0.0439

Max Act: 0.0366. Min Act: -0.0003

Data Index: 594149 (C4 (Web Text))

Max Activating Token Index: 34