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

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

Neuron 3487 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: 0.3088. Min Range: -0.3088

Max Act: 0.3088. Min Act: -0.0000

Data Index: 366294 (C4 (Web Text))

Max Activating Token Index: 895

Text #1

Max Range: 0.3088. Min Range: -0.3088

Max Act: 0.2900. Min Act: -0.0001

Data Index: 1360031 (Python Code)

Max Activating Token Index: 910

Text #2

Max Range: 0.3088. Min Range: -0.3088

Max Act: 0.2877. Min Act: -0.0001

Data Index: 1041946 (C4 (Web Text))

Max Activating Token Index: 1002

Text #3

Max Range: 0.3088. Min Range: -0.3088

Max Act: 0.2736. Min Act: -0.0001

Data Index: 648719 (C4 (Web Text))

Max Activating Token Index: 138

Text #4

Max Range: 0.3088. Min Range: -0.3088

Max Act: 0.2751. Min Act: -0.0001

Data Index: 1619771 (Python Code)

Max Activating Token Index: 484

Text #5

Max Range: 0.3088. Min Range: -0.3088

Max Act: 0.2715. Min Act: -0.0000

Data Index: 241952 (C4 (Web Text))

Max Activating Token Index: 529

Text #6

Max Range: 0.3088. Min Range: -0.3088

Max Act: 0.2576. Min Act: -0.0000

Data Index: 158816 (C4 (Web Text))

Max Activating Token Index: 274

Text #7

Max Range: 0.3088. Min Range: -0.3088

Max Act: 0.2709. Min Act: -0.0001

Data Index: 1527280 (Python Code)

Max Activating Token Index: 355

Text #8

Max Range: 0.3088. Min Range: -0.3088

Max Act: 0.2541. Min Act: -0.0001

Data Index: 519471 (C4 (Web Text))

Max Activating Token Index: 431

Text #9

Max Range: 0.3088. Min Range: -0.3088

Max Act: 0.2512. Min Act: -0.0001

Data Index: 266118 (C4 (Web Text))

Max Activating Token Index: 36

Text #10

Max Range: 0.3088. Min Range: -0.3088

Max Act: 0.2503. Min Act: -0.0001

Data Index: 1449189 (Python Code)

Max Activating Token Index: 896

Text #11

Max Range: 0.3088. Min Range: -0.3088

Max Act: 0.2508. Min Act: -0.0001

Data Index: 954972 (C4 (Web Text))

Max Activating Token Index: 985

Text #12

Max Range: 0.3088. Min Range: -0.3088

Max Act: 0.2495. Min Act: -0.0001

Data Index: 1180656 (C4 (Web Text))

Max Activating Token Index: 23

Text #13

Max Range: 0.3088. Min Range: -0.3088

Max Act: 0.2375. Min Act: -0.0000

Data Index: 380009 (C4 (Web Text))

Max Activating Token Index: 123

Text #14

Max Range: 0.3088. Min Range: -0.3088

Max Act: 0.2405. Min Act: -0.0001

Data Index: 598139 (C4 (Web Text))

Max Activating Token Index: 256

Text #15

Max Range: 0.3088. Min Range: -0.3088

Max Act: 0.2342. Min Act: -0.0000

Data Index: 175139 (C4 (Web Text))

Max Activating Token Index: 741

Text #16

Max Range: 0.3088. Min Range: -0.3088

Max Act: 0.2207. Min Act: -0.0000

Data Index: 982787 (C4 (Web Text))

Max Activating Token Index: 110

Text #17

Max Range: 0.3088. Min Range: -0.3088

Max Act: 0.2245. Min Act: -0.0001

Data Index: 13035 (C4 (Web Text))

Max Activating Token Index: 19

Text #18

Max Range: 0.3088. Min Range: -0.3088

Max Act: 0.2225. Min Act: -0.0000

Data Index: 828462 (C4 (Web Text))

Max Activating Token Index: 406

Text #19

Max Range: 0.3088. Min Range: -0.3088

Max Act: 0.2379. Min Act: -0.0000

Data Index: 1352626 (C4 (Web Text))

Max Activating Token Index: 392