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

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

Neuron 1763 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: 0.4392. Min Range: -0.4392

Max Act: 0.4379. Min Act: -0.0000

Data Index: 1170218 (C4 (Web Text))

Max Activating Token Index: 1016

Text #1

Max Range: 0.4392. Min Range: -0.4392

Max Act: 0.4392. Min Act: -0.0000

Data Index: 313121 (C4 (Web Text))

Max Activating Token Index: 543

Text #2

Max Range: 0.4392. Min Range: -0.4392

Max Act: 0.4082. Min Act: -0.0000

Data Index: 586006 (C4 (Web Text))

Max Activating Token Index: 280

Text #3

Max Range: 0.4392. Min Range: -0.4392

Max Act: 0.4144. Min Act: -0.0000

Data Index: 886058 (C4 (Web Text))

Max Activating Token Index: 37

Text #4

Max Range: 0.4392. Min Range: -0.4392

Max Act: 0.4278. Min Act: -0.0000

Data Index: 828605 (C4 (Web Text))

Max Activating Token Index: 463

Text #5

Max Range: 0.4392. Min Range: -0.4392

Max Act: 0.4129. Min Act: -0.0000

Data Index: 490446 (C4 (Web Text))

Max Activating Token Index: 576

Text #6

Max Range: 0.4392. Min Range: -0.4392

Max Act: 0.4014. Min Act: -0.0000

Data Index: 1304357 (C4 (Web Text))

Max Activating Token Index: 741

Text #7

Max Range: 0.4392. Min Range: -0.4392

Max Act: 0.3973. Min Act: -0.0000

Data Index: 824845 (C4 (Web Text))

Max Activating Token Index: 1009

Text #8

Max Range: 0.4392. Min Range: -0.4392

Max Act: 0.3940. Min Act: -0.0000

Data Index: 98186 (C4 (Web Text))

Max Activating Token Index: 368

Text #9

Max Range: 0.4392. Min Range: -0.4392

Max Act: 0.3946. Min Act: -0.0000

Data Index: 1563402 (Python Code)

Max Activating Token Index: 686

Text #10

Max Range: 0.4392. Min Range: -0.4392

Max Act: 0.3716. Min Act: -0.0000

Data Index: 389107 (C4 (Web Text))

Max Activating Token Index: 912

Text #11

Max Range: 0.4392. Min Range: -0.4392

Max Act: 0.3619. Min Act: -0.0000

Data Index: 1354378 (C4 (Web Text))

Max Activating Token Index: 626

Text #12

Max Range: 0.4392. Min Range: -0.4392

Max Act: 0.3699. Min Act: 0.0000

Data Index: 492580 (C4 (Web Text))

Max Activating Token Index: 504

Text #13

Max Range: 0.4392. Min Range: -0.4392

Max Act: 0.3584. Min Act: -0.0000

Data Index: 1388317 (Python Code)

Max Activating Token Index: 180

Text #14

Max Range: 0.4392. Min Range: -0.4392

Max Act: 0.3560. Min Act: -0.0000

Data Index: 37540 (C4 (Web Text))

Max Activating Token Index: 178

Text #15

Max Range: 0.4392. Min Range: -0.4392

Max Act: 0.3663. Min Act: -0.0000

Data Index: 14963 (C4 (Web Text))

Max Activating Token Index: 788

Text #16

Max Range: 0.4392. Min Range: -0.4392

Max Act: 0.3469. Min Act: -0.0000

Data Index: 1185928 (C4 (Web Text))

Max Activating Token Index: 476

Text #17

Max Range: 0.4392. Min Range: -0.4392

Max Act: 0.3436. Min Act: -0.0000

Data Index: 588783 (C4 (Web Text))

Max Activating Token Index: 284

Text #18

Max Range: 0.4392. Min Range: -0.4392

Max Act: 0.3545. Min Act: -0.0000

Data Index: 1061339 (C4 (Web Text))

Max Activating Token Index: 432

Text #19

Max Range: 0.4392. Min Range: -0.4392

Max Act: 0.3492. Min Act: -0.0000

Data Index: 1518224 (Python Code)

Max Activating Token Index: 381