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

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

Neuron 3256 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.6885. Min Act: -0.0001

Data Index: 584021 (C4 (Web Text))

Max Activating Token Index: 362

Text #1

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.6501. Min Act: -0.0002

Data Index: 769463 (C4 (Web Text))

Max Activating Token Index: 662

Text #2

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.6319. Min Act: -0.0002

Data Index: 703801 (C4 (Web Text))

Max Activating Token Index: 719

Text #3

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.6496. Min Act: -0.0002

Data Index: 584440 (C4 (Web Text))

Max Activating Token Index: 317

Text #4

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.6392. Min Act: -0.0001

Data Index: 149167 (C4 (Web Text))

Max Activating Token Index: 969

Text #5

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.6815. Min Act: -0.0001

Data Index: 1251026 (C4 (Web Text))

Max Activating Token Index: 390

Text #6

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.6561. Min Act: -0.0002

Data Index: 284522 (C4 (Web Text))

Max Activating Token Index: 705

Text #7

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.6568. Min Act: -0.0002

Data Index: 1115826 (C4 (Web Text))

Max Activating Token Index: 251

Text #8

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.6414. Min Act: -0.0001

Data Index: 839156 (C4 (Web Text))

Max Activating Token Index: 720

Text #9

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.6307. Min Act: -0.0002

Data Index: 459186 (C4 (Web Text))

Max Activating Token Index: 305

Text #10

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.6581. Min Act: -0.0001

Data Index: 85154 (C4 (Web Text))

Max Activating Token Index: 906

Text #11

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.6104. Min Act: -0.0001

Data Index: 904310 (C4 (Web Text))

Max Activating Token Index: 529

Text #12

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.5942. Min Act: -0.0001

Data Index: 1326697 (C4 (Web Text))

Max Activating Token Index: 862

Text #13

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.6147. Min Act: -0.0001

Data Index: 933449 (C4 (Web Text))

Max Activating Token Index: 116

Text #14

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.5779. Min Act: -0.0001

Data Index: 44788 (C4 (Web Text))

Max Activating Token Index: 218

Text #15

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.5930. Min Act: -0.0002

Data Index: 1357231 (C4 (Web Text))

Max Activating Token Index: 984

Text #16

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.5962. Min Act: -0.0002

Data Index: 882003 (C4 (Web Text))

Max Activating Token Index: 823

Text #17

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.5919. Min Act: -0.0002

Data Index: 1107280 (C4 (Web Text))

Max Activating Token Index: 151

Text #18

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.6344. Min Act: -0.0002

Data Index: 1097405 (C4 (Web Text))

Max Activating Token Index: 750

Text #19

Max Range: 0.6885. Min Range: -0.6885

Max Act: 0.5349. Min Act: -0.0002

Data Index: 1175242 (C4 (Web Text))

Max Activating Token Index: 952