Model: SoLU Model: 8 Layers, 4096 Neurons per Layer

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

Neuron 3072 in Layer 6

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.3920. Min Act: -0.0001

Data Index: 255734 (C4 (Web Text))

Max Activating Token Index: 858

Text #1

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.4138. Min Act: -0.0001

Data Index: 1001358 (C4 (Web Text))

Max Activating Token Index: 528

Text #2

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.4482. Min Act: -0.0001

Data Index: 1291796 (C4 (Web Text))

Max Activating Token Index: 635

Text #3

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.3627. Min Act: -0.0000

Data Index: 172860 (C4 (Web Text))

Max Activating Token Index: 564

Text #4

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.3537. Min Act: -0.0001

Data Index: 1085352 (C4 (Web Text))

Max Activating Token Index: 509

Text #5

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.3051. Min Act: -0.0001

Data Index: 583154 (C4 (Web Text))

Max Activating Token Index: 656

Text #6

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.3104. Min Act: -0.0001

Data Index: 122267 (C4 (Web Text))

Max Activating Token Index: 576

Text #7

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.3142. Min Act: -0.0000

Data Index: 1263769 (C4 (Web Text))

Max Activating Token Index: 196

Text #8

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.3038. Min Act: -0.0001

Data Index: 820378 (C4 (Web Text))

Max Activating Token Index: 722

Text #9

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.3109. Min Act: -0.0001

Data Index: 1126046 (C4 (Web Text))

Max Activating Token Index: 500

Text #10

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.3186. Min Act: -0.0001

Data Index: 536072 (C4 (Web Text))

Max Activating Token Index: 949

Text #11

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.3029. Min Act: -0.0001

Data Index: 422869 (C4 (Web Text))

Max Activating Token Index: 66

Text #12

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.3047. Min Act: -0.0001

Data Index: 1121542 (C4 (Web Text))

Max Activating Token Index: 928

Text #13

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.2853. Min Act: -0.0000

Data Index: 1173993 (C4 (Web Text))

Max Activating Token Index: 173

Text #14

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.2881. Min Act: -0.0000

Data Index: 289456 (C4 (Web Text))

Max Activating Token Index: 54

Text #15

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.2645. Min Act: -0.0001

Data Index: 125216 (C4 (Web Text))

Max Activating Token Index: 375

Text #16

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.2723. Min Act: -0.0000

Data Index: 297193 (C4 (Web Text))

Max Activating Token Index: 244

Text #17

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.2579. Min Act: -0.0000

Data Index: 1176180 (C4 (Web Text))

Max Activating Token Index: 22

Text #18

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.2572. Min Act: -0.0001

Data Index: 333644 (C4 (Web Text))

Max Activating Token Index: 1001

Text #19

Max Range: 0.4482. Min Range: -0.4482

Max Act: 0.2671. Min Act: -0.0000

Data Index: 81381 (C4 (Web Text))

Max Activating Token Index: 177