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

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

Neuron 455 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: 0.3810. Min Range: -0.3810

Max Act: 0.3810. Min Act: -0.0000

Data Index: 812038 (C4 (Web Text))

Max Activating Token Index: 461

Text #1

Max Range: 0.3810. Min Range: -0.3810

Max Act: 0.3375. Min Act: -0.0000

Data Index: 869780 (C4 (Web Text))

Max Activating Token Index: 752

Text #2

Max Range: 0.3810. Min Range: -0.3810

Max Act: 0.2844. Min Act: -0.0000

Data Index: 1108563 (C4 (Web Text))

Max Activating Token Index: 685

Text #3

Max Range: 0.3810. Min Range: -0.3810

Max Act: 0.2482. Min Act: -0.0000

Data Index: 1351745 (C4 (Web Text))

Max Activating Token Index: 969

Text #4

Max Range: 0.3810. Min Range: -0.3810

Max Act: 0.2261. Min Act: -0.0000

Data Index: 666369 (C4 (Web Text))

Max Activating Token Index: 355

Text #5

Max Range: 0.3810. Min Range: -0.3810

Max Act: 0.2344. Min Act: -0.0000

Data Index: 623970 (C4 (Web Text))

Max Activating Token Index: 231

Text #6

Max Range: 0.3810. Min Range: -0.3810

Max Act: 0.2359. Min Act: -0.0000

Data Index: 1621406 (Python Code)

Max Activating Token Index: 803

Text #7

Max Range: 0.3810. Min Range: -0.3810

Max Act: 0.2476. Min Act: -0.0000

Data Index: 634384 (C4 (Web Text))

Max Activating Token Index: 626

Text #8

Max Range: 0.3810. Min Range: -0.3810

Max Act: 0.2374. Min Act: 0.0000

Data Index: 1062668 (C4 (Web Text))

Max Activating Token Index: 429

Text #9

Max Range: 0.3810. Min Range: -0.3810

Max Act: 0.2448. Min Act: -0.0000

Data Index: 1656760 (Python Code)

Max Activating Token Index: 637

Text #10

Max Range: 0.3810. Min Range: -0.3810

Max Act: 0.2352. Min Act: -0.0000

Data Index: 571336 (C4 (Web Text))

Max Activating Token Index: 618

Text #11

Max Range: 0.3810. Min Range: -0.3810

Max Act: 0.2303. Min Act: -0.0000

Data Index: 7935 (C4 (Web Text))

Max Activating Token Index: 481

Text #12

Max Range: 0.3810. Min Range: -0.3810

Max Act: 0.2382. Min Act: -0.0000

Data Index: 139813 (C4 (Web Text))

Max Activating Token Index: 614

Text #13

Max Range: 0.3810. Min Range: -0.3810

Max Act: 0.2207. Min Act: -0.0000

Data Index: 614502 (C4 (Web Text))

Max Activating Token Index: 812

Text #14

Max Range: 0.3810. Min Range: -0.3810

Max Act: 0.2197. Min Act: -0.0000

Data Index: 296677 (C4 (Web Text))

Max Activating Token Index: 66

Text #15

Max Range: 0.3810. Min Range: -0.3810

Max Act: 0.2113. Min Act: -0.0000

Data Index: 1270584 (C4 (Web Text))

Max Activating Token Index: 652

Text #16

Max Range: 0.3810. Min Range: -0.3810

Max Act: 0.2086. Min Act: -0.0000

Data Index: 1575141 (Python Code)

Max Activating Token Index: 783

Text #17

Max Range: 0.3810. Min Range: -0.3810

Max Act: 0.2243. Min Act: -0.0000

Data Index: 893606 (C4 (Web Text))

Max Activating Token Index: 691

Text #18

Max Range: 0.3810. Min Range: -0.3810

Max Act: 0.2127. Min Act: -0.0000

Data Index: 607185 (C4 (Web Text))

Max Activating Token Index: 372

Text #19

Max Range: 0.3810. Min Range: -0.3810

Max Act: 0.2123. Min Act: -0.0000

Data Index: 633389 (C4 (Web Text))

Max Activating Token Index: 27