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

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

Neuron 4720 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: 2.3881. Min Range: -2.3881

Max Act: 2.3881. Min Act: -0.0000

Data Index: 581923 (C4 (Web Text))

Max Activating Token Index: 95

Text #1

Max Range: 2.3881. Min Range: -2.3881

Max Act: 2.2513. Min Act: -0.0000

Data Index: 1260245 (C4 (Web Text))

Max Activating Token Index: 76

Text #2

Max Range: 2.3881. Min Range: -2.3881

Max Act: 2.2186. Min Act: -0.0000

Data Index: 214303 (C4 (Web Text))

Max Activating Token Index: 887

Text #3

Max Range: 2.3881. Min Range: -2.3881

Max Act: 2.2043. Min Act: -0.0000

Data Index: 214376 (C4 (Web Text))

Max Activating Token Index: 99

Text #4

Max Range: 2.3881. Min Range: -2.3881

Max Act: 2.1493. Min Act: -0.0000

Data Index: 519246 (C4 (Web Text))

Max Activating Token Index: 758

Text #5

Max Range: 2.3881. Min Range: -2.3881

Max Act: 1.9000. Min Act: -0.0000

Data Index: 61284 (C4 (Web Text))

Max Activating Token Index: 180

Text #6

Max Range: 2.3881. Min Range: -2.3881

Max Act: 2.2053. Min Act: -0.0000

Data Index: 986873 (C4 (Web Text))

Max Activating Token Index: 24

Text #7

Max Range: 2.3881. Min Range: -2.3881

Max Act: 1.9895. Min Act: -0.0000

Data Index: 112211 (C4 (Web Text))

Max Activating Token Index: 912

Text #8

Max Range: 2.3881. Min Range: -2.3881

Max Act: 2.1042. Min Act: -0.0000

Data Index: 319281 (C4 (Web Text))

Max Activating Token Index: 72

Text #9

Max Range: 2.3881. Min Range: -2.3881

Max Act: 1.9837. Min Act: -0.0000

Data Index: 330234 (C4 (Web Text))

Max Activating Token Index: 854

Text #10

Max Range: 2.3881. Min Range: -2.3881

Max Act: 1.9702. Min Act: -0.0000

Data Index: 888253 (C4 (Web Text))

Max Activating Token Index: 494

Text #11

Max Range: 2.3881. Min Range: -2.3881

Max Act: 1.8805. Min Act: -0.0000

Data Index: 1161516 (C4 (Web Text))

Max Activating Token Index: 445

Text #12

Max Range: 2.3881. Min Range: -2.3881

Max Act: 1.8522. Min Act: -0.0000

Data Index: 139589 (C4 (Web Text))

Max Activating Token Index: 83

Text #13

Max Range: 2.3881. Min Range: -2.3881

Max Act: 2.0484. Min Act: -0.0000

Data Index: 310996 (C4 (Web Text))

Max Activating Token Index: 647

Text #14

Max Range: 2.3881. Min Range: -2.3881

Max Act: 2.1448. Min Act: -0.0000

Data Index: 1416546 (Python Code)

Max Activating Token Index: 839

Text #15

Max Range: 2.3881. Min Range: -2.3881

Max Act: 1.9000. Min Act: -0.0000

Data Index: 637330 (C4 (Web Text))

Max Activating Token Index: 462

Text #16

Max Range: 2.3881. Min Range: -2.3881

Max Act: 1.8877. Min Act: -0.0000

Data Index: 204683 (C4 (Web Text))

Max Activating Token Index: 184

Text #17

Max Range: 2.3881. Min Range: -2.3881

Max Act: 1.8790. Min Act: -0.0000

Data Index: 1354448 (C4 (Web Text))

Max Activating Token Index: 447

Text #18

Max Range: 2.3881. Min Range: -2.3881

Max Act: 1.8449. Min Act: -0.0000

Data Index: 1379858 (Python Code)

Max Activating Token Index: 11

Text #19

Max Range: 2.3881. Min Range: -2.3881

Max Act: 1.8475. Min Act: -0.0000

Data Index: 993337 (C4 (Web Text))

Max Activating Token Index: 714