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

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

Neuron 908 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0640. Min Act: -0.0000

Data Index: 500639 (C4 (Web Text))

Max Activating Token Index: 146

Text #1

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0612. Min Act: -0.0000

Data Index: 892551 (C4 (Web Text))

Max Activating Token Index: 421

Text #2

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0621. Min Act: -0.0000

Data Index: 1114518 (C4 (Web Text))

Max Activating Token Index: 220

Text #3

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0581. Min Act: -0.0000

Data Index: 221396 (C4 (Web Text))

Max Activating Token Index: 227

Text #4

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0648. Min Act: -0.0000

Data Index: 960115 (C4 (Web Text))

Max Activating Token Index: 613

Text #5

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0595. Min Act: -0.0001

Data Index: 1207737 (C4 (Web Text))

Max Activating Token Index: 975

Text #6

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0584. Min Act: -0.0000

Data Index: 468511 (C4 (Web Text))

Max Activating Token Index: 123

Text #7

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0587. Min Act: -0.0000

Data Index: 466691 (C4 (Web Text))

Max Activating Token Index: 256

Text #8

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0575. Min Act: -0.0000

Data Index: 1230188 (C4 (Web Text))

Max Activating Token Index: 512

Text #9

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0563. Min Act: -0.0000

Data Index: 415128 (C4 (Web Text))

Max Activating Token Index: 882

Text #10

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0589. Min Act: -0.0000

Data Index: 326862 (C4 (Web Text))

Max Activating Token Index: 970

Text #11

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0544. Min Act: -0.0000

Data Index: 598716 (C4 (Web Text))

Max Activating Token Index: 533

Text #12

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0545. Min Act: -0.0001

Data Index: 1413425 (Python Code)

Max Activating Token Index: 837

Text #13

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0549. Min Act: -0.0000

Data Index: 800562 (C4 (Web Text))

Max Activating Token Index: 236

Text #14

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0544. Min Act: -0.0000

Data Index: 1141019 (C4 (Web Text))

Max Activating Token Index: 482

Text #15

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0559. Min Act: -0.0001

Data Index: 1319421 (C4 (Web Text))

Max Activating Token Index: 567

Text #16

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0507. Min Act: -0.0000

Data Index: 456874 (C4 (Web Text))

Max Activating Token Index: 961

Text #17

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0535. Min Act: -0.0001

Data Index: 973282 (C4 (Web Text))

Max Activating Token Index: 864

Text #18

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0535. Min Act: -0.0001

Data Index: 392346 (C4 (Web Text))

Max Activating Token Index: 321

Text #19

Max Range: 0.0648. Min Range: -0.0648

Max Act: 0.0506. Min Act: -0.0000

Data Index: 123875 (C4 (Web Text))

Max Activating Token Index: 914