Model: SoLU Model: 12 Layers, 6144 Neurons per Layer

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

Neuron 3798 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-12l')

Text #0

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.1214. Min Act: -0.0000

Data Index: 1017912 (C4 (Web Text))

Max Activating Token Index: 836

Text #1

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.1215. Min Act: -0.0000

Data Index: 1398830 (Python Code)

Max Activating Token Index: 13

Text #2

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.1159. Min Act: -0.0000

Data Index: 1584511 (Python Code)

Max Activating Token Index: 13

Text #3

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.1069. Min Act: -0.0000

Data Index: 1475691 (Python Code)

Max Activating Token Index: 307

Text #4

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.0959. Min Act: -0.0000

Data Index: 710026 (C4 (Web Text))

Max Activating Token Index: 678

Text #5

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.0986. Min Act: -0.0000

Data Index: 461737 (C4 (Web Text))

Max Activating Token Index: 888

Text #6

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.1068. Min Act: -0.0001

Data Index: 794971 (C4 (Web Text))

Max Activating Token Index: 518

Text #7

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.0982. Min Act: -0.0000

Data Index: 1154408 (C4 (Web Text))

Max Activating Token Index: 278

Text #8

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.0860. Min Act: -0.0000

Data Index: 895598 (C4 (Web Text))

Max Activating Token Index: 668

Text #9

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.0996. Min Act: -0.0000

Data Index: 578580 (C4 (Web Text))

Max Activating Token Index: 928

Text #10

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.1002. Min Act: -0.0000

Data Index: 1563872 (Python Code)

Max Activating Token Index: 29

Text #11

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.0896. Min Act: -0.0000

Data Index: 1397655 (Python Code)

Max Activating Token Index: 479

Text #12

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.0928. Min Act: -0.0001

Data Index: 1232420 (C4 (Web Text))

Max Activating Token Index: 393

Text #13

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.0866. Min Act: -0.0001

Data Index: 926681 (C4 (Web Text))

Max Activating Token Index: 743

Text #14

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.0923. Min Act: -0.0000

Data Index: 1458695 (Python Code)

Max Activating Token Index: 485

Text #15

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.0931. Min Act: -0.0000

Data Index: 1158431 (C4 (Web Text))

Max Activating Token Index: 934

Text #16

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.0956. Min Act: -0.0000

Data Index: 1068941 (C4 (Web Text))

Max Activating Token Index: 808

Text #17

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.1026. Min Act: -0.0001

Data Index: 501181 (C4 (Web Text))

Max Activating Token Index: 885

Text #18

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.0872. Min Act: -0.0001

Data Index: 1033433 (C4 (Web Text))

Max Activating Token Index: 409

Text #19

Max Range: 0.1215. Min Range: -0.1215

Max Act: 0.0907. Min Act: -0.0000

Data Index: 1110327 (C4 (Web Text))

Max Activating Token Index: 223