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

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

Neuron 6 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: 0.4857. Min Range: -0.4857

Max Act: 0.4857. Min Act: -0.0000

Data Index: 1641450 (Python Code)

Max Activating Token Index: 784

Text #1

Max Range: 0.4857. Min Range: -0.4857

Max Act: 0.4323. Min Act: -0.0000

Data Index: 1372294 (Python Code)

Max Activating Token Index: 481

Text #2

Max Range: 0.4857. Min Range: -0.4857

Max Act: 0.3896. Min Act: -0.0000

Data Index: 764513 (C4 (Web Text))

Max Activating Token Index: 722

Text #3

Max Range: 0.4857. Min Range: -0.4857

Max Act: 0.3696. Min Act: -0.0000

Data Index: 667141 (C4 (Web Text))

Max Activating Token Index: 569

Text #4

Max Range: 0.4857. Min Range: -0.4857

Max Act: 0.3671. Min Act: -0.0000

Data Index: 1177148 (C4 (Web Text))

Max Activating Token Index: 295

Text #5

Max Range: 0.4857. Min Range: -0.4857

Max Act: 0.3913. Min Act: -0.0000

Data Index: 714413 (C4 (Web Text))

Max Activating Token Index: 342

Text #6

Max Range: 0.4857. Min Range: -0.4857

Max Act: 0.3663. Min Act: -0.0000

Data Index: 250081 (C4 (Web Text))

Max Activating Token Index: 760

Text #7

Max Range: 0.4857. Min Range: -0.4857

Max Act: 0.3698. Min Act: -0.0000

Data Index: 1222782 (C4 (Web Text))

Max Activating Token Index: 736

Text #8

Max Range: 0.4857. Min Range: -0.4857

Max Act: 0.3648. Min Act: -0.0000

Data Index: 207959 (C4 (Web Text))

Max Activating Token Index: 170

Text #9

Max Range: 0.4857. Min Range: -0.4857

Max Act: 0.3193. Min Act: -0.0000

Data Index: 532251 (C4 (Web Text))

Max Activating Token Index: 562

Text #10

Max Range: 0.4857. Min Range: -0.4857

Max Act: 0.3254. Min Act: -0.0000

Data Index: 1373554 (Python Code)

Max Activating Token Index: 63

Text #11

Max Range: 0.4857. Min Range: -0.4857

Max Act: 0.3018. Min Act: -0.0000

Data Index: 1631601 (Python Code)

Max Activating Token Index: 925

Text #12

Max Range: 0.4857. Min Range: -0.4857

Max Act: 0.2987. Min Act: -0.0000

Data Index: 1138356 (C4 (Web Text))

Max Activating Token Index: 892

Text #13

Max Range: 0.4857. Min Range: -0.4857

Max Act: 0.3136. Min Act: -0.0000

Data Index: 1559489 (Python Code)

Max Activating Token Index: 129

Text #14

Max Range: 0.4857. Min Range: -0.4857

Max Act: 0.3197. Min Act: -0.0000

Data Index: 1408517 (Python Code)

Max Activating Token Index: 220

Text #15

Max Range: 0.4857. Min Range: -0.4857

Max Act: 0.3147. Min Act: -0.0000

Data Index: 162342 (C4 (Web Text))

Max Activating Token Index: 814

Text #16

Max Range: 0.4857. Min Range: -0.4857

Max Act: 0.3232. Min Act: -0.0000

Data Index: 699814 (C4 (Web Text))

Max Activating Token Index: 892

Text #17

Max Range: 0.4857. Min Range: -0.4857

Max Act: 0.2863. Min Act: -0.0000

Data Index: 1605230 (Python Code)

Max Activating Token Index: 651

Text #18

Max Range: 0.4857. Min Range: -0.4857

Max Act: 0.2825. Min Act: -0.0000

Data Index: 917338 (C4 (Web Text))

Max Activating Token Index: 846

Text #19

Max Range: 0.4857. Min Range: -0.4857

Max Act: 0.2912. Min Act: -0.0000

Data Index: 353813 (C4 (Web Text))

Max Activating Token Index: 707