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

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

Neuron 4921 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: 0.8914. Min Range: -0.8914

Max Act: 0.8914. Min Act: -0.0000

Data Index: 325572 (C4 (Web Text))

Max Activating Token Index: 80

Text #1

Max Range: 0.8914. Min Range: -0.8914

Max Act: 0.8520. Min Act: -0.0000

Data Index: 577717 (C4 (Web Text))

Max Activating Token Index: 359

Text #2

Max Range: 0.8914. Min Range: -0.8914

Max Act: 0.8640. Min Act: -0.0001

Data Index: 116996 (C4 (Web Text))

Max Activating Token Index: 114

Text #3

Max Range: 0.8914. Min Range: -0.8914

Max Act: 0.8072. Min Act: -0.0000

Data Index: 521132 (C4 (Web Text))

Max Activating Token Index: 4

Text #4

Max Range: 0.8914. Min Range: -0.8914

Max Act: 0.8280. Min Act: -0.0000

Data Index: 519860 (C4 (Web Text))

Max Activating Token Index: 514

Text #5

Max Range: 0.8914. Min Range: -0.8914

Max Act: 0.8391. Min Act: -0.0000

Data Index: 174283 (C4 (Web Text))

Max Activating Token Index: 839

Text #6

Max Range: 0.8914. Min Range: -0.8914

Max Act: 0.7987. Min Act: -0.0001

Data Index: 592960 (C4 (Web Text))

Max Activating Token Index: 990

Text #7

Max Range: 0.8914. Min Range: -0.8914

Max Act: 0.7954. Min Act: -0.0001

Data Index: 58389 (C4 (Web Text))

Max Activating Token Index: 236

Text #8

Max Range: 0.8914. Min Range: -0.8914

Max Act: 0.7684. Min Act: -0.0000

Data Index: 71502 (C4 (Web Text))

Max Activating Token Index: 274

Text #9

Max Range: 0.8914. Min Range: -0.8914

Max Act: 0.7546. Min Act: -0.0001

Data Index: 102951 (C4 (Web Text))

Max Activating Token Index: 924

Text #10

Max Range: 0.8914. Min Range: -0.8914

Max Act: 0.8158. Min Act: -0.0000

Data Index: 1357830 (C4 (Web Text))

Max Activating Token Index: 353

Text #11

Max Range: 0.8914. Min Range: -0.8914

Max Act: 0.6675. Min Act: -0.0000

Data Index: 76727 (C4 (Web Text))

Max Activating Token Index: 780

Text #12

Max Range: 0.8914. Min Range: -0.8914

Max Act: 0.7070. Min Act: -0.0001

Data Index: 529347 (C4 (Web Text))

Max Activating Token Index: 866

Text #13

Max Range: 0.8914. Min Range: -0.8914

Max Act: 0.7315. Min Act: -0.0000

Data Index: 858160 (C4 (Web Text))

Max Activating Token Index: 220

Text #14

Max Range: 0.8914. Min Range: -0.8914

Max Act: 0.6832. Min Act: -0.0000

Data Index: 148043 (C4 (Web Text))

Max Activating Token Index: 507

Text #15

Max Range: 0.8914. Min Range: -0.8914

Max Act: 0.7257. Min Act: -0.0000

Data Index: 1266830 (C4 (Web Text))

Max Activating Token Index: 428

Text #16

Max Range: 0.8914. Min Range: -0.8914

Max Act: 0.6891. Min Act: -0.0000

Data Index: 1412158 (Python Code)

Max Activating Token Index: 91

Text #17

Max Range: 0.8914. Min Range: -0.8914

Max Act: 0.7116. Min Act: -0.0001

Data Index: 506395 (C4 (Web Text))

Max Activating Token Index: 667

Text #18

Max Range: 0.8914. Min Range: -0.8914

Max Act: 0.7199. Min Act: -0.0000

Data Index: 1620632 (Python Code)

Max Activating Token Index: 367

Text #19

Max Range: 0.8914. Min Range: -0.8914

Max Act: 0.7164. Min Act: -0.0000

Data Index: 1041120 (C4 (Web Text))

Max Activating Token Index: 146