Model: SoLU Model: 6 Layers, 3072 Neurons per Layer

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

Neuron 552 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-6l')

Text #0

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.6975. Min Act: -0.0001

Data Index: 41389 (C4 (Web Text))

Max Activating Token Index: 277

Text #1

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.5818. Min Act: -0.0001

Data Index: 720108 (C4 (Web Text))

Max Activating Token Index: 817

Text #2

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.6297. Min Act: -0.0001

Data Index: 1574972 (Python Code)

Max Activating Token Index: 273

Text #3

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.5861. Min Act: -0.0001

Data Index: 438838 (C4 (Web Text))

Max Activating Token Index: 127

Text #4

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.5686. Min Act: -0.0001

Data Index: 13130 (C4 (Web Text))

Max Activating Token Index: 200

Text #5

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.5096. Min Act: -0.0001

Data Index: 720414 (C4 (Web Text))

Max Activating Token Index: 518

Text #6

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.5134. Min Act: -0.0001

Data Index: 1483643 (Python Code)

Max Activating Token Index: 458

Text #7

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.4865. Min Act: -0.0001

Data Index: 1530973 (Python Code)

Max Activating Token Index: 563

Text #8

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.5329. Min Act: -0.0001

Data Index: 446417 (C4 (Web Text))

Max Activating Token Index: 890

Text #9

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.5535. Min Act: -0.0001

Data Index: 911232 (C4 (Web Text))

Max Activating Token Index: 298

Text #10

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.4873. Min Act: -0.0001

Data Index: 650737 (C4 (Web Text))

Max Activating Token Index: 406

Text #11

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.4670. Min Act: -0.0001

Data Index: 615365 (C4 (Web Text))

Max Activating Token Index: 109

Text #12

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.4365. Min Act: -0.0001

Data Index: 464423 (C4 (Web Text))

Max Activating Token Index: 265

Text #13

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.4827. Min Act: -0.0001

Data Index: 1464065 (Python Code)

Max Activating Token Index: 772

Text #14

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.4677. Min Act: -0.0001

Data Index: 409139 (C4 (Web Text))

Max Activating Token Index: 537

Text #15

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.4584. Min Act: -0.0000

Data Index: 447378 (C4 (Web Text))

Max Activating Token Index: 650

Text #16

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.4455. Min Act: -0.0001

Data Index: 598906 (C4 (Web Text))

Max Activating Token Index: 286

Text #17

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.4630. Min Act: -0.0001

Data Index: 234169 (C4 (Web Text))

Max Activating Token Index: 772

Text #18

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.4682. Min Act: -0.0001

Data Index: 226476 (C4 (Web Text))

Max Activating Token Index: 208

Text #19

Max Range: 0.6975. Min Range: -0.6975

Max Act: 0.4500. Min Act: -0.0001

Data Index: 791022 (C4 (Web Text))

Max Activating Token Index: 494