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

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

Neuron 58 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: 0.2872. Min Range: -0.2872

Max Act: 0.2872. Min Act: -0.0002

Data Index: 704232 (C4 (Web Text))

Max Activating Token Index: 938

Text #1

Max Range: 0.2872. Min Range: -0.2872

Max Act: 0.2564. Min Act: -0.0002

Data Index: 872661 (C4 (Web Text))

Max Activating Token Index: 876

Text #2

Max Range: 0.2872. Min Range: -0.2872

Max Act: 0.2604. Min Act: -0.0002

Data Index: 316178 (C4 (Web Text))

Max Activating Token Index: 794

Text #3

Max Range: 0.2872. Min Range: -0.2872

Max Act: 0.2517. Min Act: -0.0002

Data Index: 266703 (C4 (Web Text))

Max Activating Token Index: 754

Text #4

Max Range: 0.2872. Min Range: -0.2872

Max Act: 0.2547. Min Act: -0.0002

Data Index: 755964 (C4 (Web Text))

Max Activating Token Index: 629

Text #5

Max Range: 0.2872. Min Range: -0.2872

Max Act: 0.2414. Min Act: -0.0002

Data Index: 780384 (C4 (Web Text))

Max Activating Token Index: 354

Text #6

Max Range: 0.2872. Min Range: -0.2872

Max Act: 0.2478. Min Act: -0.0002

Data Index: 1073145 (C4 (Web Text))

Max Activating Token Index: 158

Text #7

Max Range: 0.2872. Min Range: -0.2872

Max Act: 0.2452. Min Act: -0.0002

Data Index: 19292 (C4 (Web Text))

Max Activating Token Index: 965

Text #8

Max Range: 0.2872. Min Range: -0.2872

Max Act: 0.2373. Min Act: -0.0002

Data Index: 114620 (C4 (Web Text))

Max Activating Token Index: 638

Text #9

Max Range: 0.2872. Min Range: -0.2872

Max Act: 0.2295. Min Act: -0.0002

Data Index: 1606729 (Python Code)

Max Activating Token Index: 225

Text #10

Max Range: 0.2872. Min Range: -0.2872

Max Act: 0.2274. Min Act: -0.0002

Data Index: 163922 (C4 (Web Text))

Max Activating Token Index: 757

Text #11

Max Range: 0.2872. Min Range: -0.2872

Max Act: 0.2268. Min Act: -0.0002

Data Index: 996252 (C4 (Web Text))

Max Activating Token Index: 187

Text #12

Max Range: 0.2872. Min Range: -0.2872

Max Act: 0.2349. Min Act: -0.0002

Data Index: 948486 (C4 (Web Text))

Max Activating Token Index: 341

Text #13

Max Range: 0.2872. Min Range: -0.2872

Max Act: 0.2192. Min Act: -0.0002

Data Index: 1293801 (C4 (Web Text))

Max Activating Token Index: 791

Text #14

Max Range: 0.2872. Min Range: -0.2872

Max Act: 0.2196. Min Act: -0.0002

Data Index: 765462 (C4 (Web Text))

Max Activating Token Index: 95

Text #15

Max Range: 0.2872. Min Range: -0.2872

Max Act: 0.2149. Min Act: -0.0002

Data Index: 770658 (C4 (Web Text))

Max Activating Token Index: 1000

Text #16

Max Range: 0.2872. Min Range: -0.2872

Max Act: 0.2306. Min Act: -0.0002

Data Index: 21274 (C4 (Web Text))

Max Activating Token Index: 812

Text #17

Max Range: 0.2872. Min Range: -0.2872

Max Act: 0.2230. Min Act: -0.0002

Data Index: 680570 (C4 (Web Text))

Max Activating Token Index: 417

Text #18

Max Range: 0.2872. Min Range: -0.2872

Max Act: 0.2177. Min Act: -0.0002

Data Index: 978291 (C4 (Web Text))

Max Activating Token Index: 1005

Text #19

Max Range: 0.2872. Min Range: -0.2872

Max Act: 0.2297. Min Act: -0.0002

Data Index: 1239484 (C4 (Web Text))

Max Activating Token Index: 765