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

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

Neuron 3500 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.6137. Min Act: -0.0000

Data Index: 1053994 (C4 (Web Text))

Max Activating Token Index: 64

Text #1

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.5020. Min Act: -0.0000

Data Index: 1110973 (C4 (Web Text))

Max Activating Token Index: 938

Text #2

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.5169. Min Act: -0.0000

Data Index: 380722 (C4 (Web Text))

Max Activating Token Index: 483

Text #3

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.5007. Min Act: -0.0000

Data Index: 1083666 (C4 (Web Text))

Max Activating Token Index: 137

Text #4

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.4620. Min Act: -0.0000

Data Index: 380793 (C4 (Web Text))

Max Activating Token Index: 511

Text #5

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.4251. Min Act: -0.0000

Data Index: 1603936 (Python Code)

Max Activating Token Index: 931

Text #6

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.4231. Min Act: -0.0000

Data Index: 1451557 (Python Code)

Max Activating Token Index: 700

Text #7

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.4224. Min Act: -0.0000

Data Index: 845469 (C4 (Web Text))

Max Activating Token Index: 533

Text #8

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.3930. Min Act: -0.0000

Data Index: 1542970 (Python Code)

Max Activating Token Index: 370

Text #9

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.3881. Min Act: -0.0000

Data Index: 758922 (C4 (Web Text))

Max Activating Token Index: 282

Text #10

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.4295. Min Act: -0.0000

Data Index: 749363 (C4 (Web Text))

Max Activating Token Index: 22

Text #11

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.4049. Min Act: -0.0000

Data Index: 140341 (C4 (Web Text))

Max Activating Token Index: 1019

Text #12

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.4077. Min Act: -0.0000

Data Index: 1346663 (C4 (Web Text))

Max Activating Token Index: 474

Text #13

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.3916. Min Act: -0.0000

Data Index: 164707 (C4 (Web Text))

Max Activating Token Index: 19

Text #14

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.3628. Min Act: -0.0000

Data Index: 1119940 (C4 (Web Text))

Max Activating Token Index: 947

Text #15

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.3864. Min Act: -0.0000

Data Index: 1561736 (Python Code)

Max Activating Token Index: 706

Text #16

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.3833. Min Act: -0.0000

Data Index: 1449728 (Python Code)

Max Activating Token Index: 28

Text #17

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.3817. Min Act: -0.0000

Data Index: 1497738 (Python Code)

Max Activating Token Index: 536

Text #18

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.3740. Min Act: -0.0000

Data Index: 865623 (C4 (Web Text))

Max Activating Token Index: 977

Text #19

Max Range: 0.6137. Min Range: -0.6137

Max Act: 0.3588. Min Act: -0.0000

Data Index: 1482050 (Python Code)

Max Activating Token Index: 282