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

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

Neuron 3232 in Layer 8

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.3509. Min Act: -0.0000

Data Index: 1586250 (Python Code)

Max Activating Token Index: 535

Text #1

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.3464. Min Act: -0.0000

Data Index: 1234402 (C4 (Web Text))

Max Activating Token Index: 884

Text #2

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.3320. Min Act: -0.0000

Data Index: 257813 (C4 (Web Text))

Max Activating Token Index: 715

Text #3

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.3133. Min Act: -0.0000

Data Index: 1567561 (Python Code)

Max Activating Token Index: 853

Text #4

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.3196. Min Act: -0.0000

Data Index: 745185 (C4 (Web Text))

Max Activating Token Index: 787

Text #5

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.2947. Min Act: -0.0000

Data Index: 46932 (C4 (Web Text))

Max Activating Token Index: 898

Text #6

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.3239. Min Act: -0.0000

Data Index: 356840 (C4 (Web Text))

Max Activating Token Index: 357

Text #7

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.3126. Min Act: -0.0000

Data Index: 634444 (C4 (Web Text))

Max Activating Token Index: 492

Text #8

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.3032. Min Act: -0.0000

Data Index: 724277 (C4 (Web Text))

Max Activating Token Index: 218

Text #9

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.3127. Min Act: 0.0000

Data Index: 317712 (C4 (Web Text))

Max Activating Token Index: 60

Text #10

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.3079. Min Act: 0.0000

Data Index: 1205603 (C4 (Web Text))

Max Activating Token Index: 217

Text #11

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.3099. Min Act: -0.0000

Data Index: 473139 (C4 (Web Text))

Max Activating Token Index: 239

Text #12

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.3093. Min Act: -0.0000

Data Index: 1568283 (Python Code)

Max Activating Token Index: 455

Text #13

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.3060. Min Act: -0.0000

Data Index: 423181 (C4 (Web Text))

Max Activating Token Index: 641

Text #14

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.3048. Min Act: -0.0000

Data Index: 388683 (C4 (Web Text))

Max Activating Token Index: 445

Text #15

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.3134. Min Act: -0.0000

Data Index: 868001 (C4 (Web Text))

Max Activating Token Index: 334

Text #16

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.3031. Min Act: 0.0001

Data Index: 469017 (C4 (Web Text))

Max Activating Token Index: 876

Text #17

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.2964. Min Act: -0.0000

Data Index: 1022737 (C4 (Web Text))

Max Activating Token Index: 766

Text #18

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.3119. Min Act: -0.0000

Data Index: 1268177 (C4 (Web Text))

Max Activating Token Index: 202

Text #19

Max Range: 0.3509. Min Range: -0.3509

Max Act: 0.2990. Min Act: -0.0000

Data Index: 839474 (C4 (Web Text))

Max Activating Token Index: 147