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

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

Neuron 2850 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: 0.6365. Min Range: -0.6365

Max Act: 0.6365. Min Act: -0.0001

Data Index: 111005 (C4 (Web Text))

Max Activating Token Index: 976

Text #1

Max Range: 0.6365. Min Range: -0.6365

Max Act: 0.5635. Min Act: -0.0000

Data Index: 160180 (C4 (Web Text))

Max Activating Token Index: 238

Text #2

Max Range: 0.6365. Min Range: -0.6365

Max Act: 0.5406. Min Act: -0.0001

Data Index: 581653 (C4 (Web Text))

Max Activating Token Index: 423

Text #3

Max Range: 0.6365. Min Range: -0.6365

Max Act: 0.5492. Min Act: -0.0000

Data Index: 936461 (C4 (Web Text))

Max Activating Token Index: 746

Text #4

Max Range: 0.6365. Min Range: -0.6365

Max Act: 0.5292. Min Act: -0.0000

Data Index: 437861 (C4 (Web Text))

Max Activating Token Index: 61

Text #5

Max Range: 0.6365. Min Range: -0.6365

Max Act: 0.5416. Min Act: -0.0000

Data Index: 775482 (C4 (Web Text))

Max Activating Token Index: 667

Text #6

Max Range: 0.6365. Min Range: -0.6365

Max Act: 0.5622. Min Act: -0.0001

Data Index: 996763 (C4 (Web Text))

Max Activating Token Index: 997

Text #7

Max Range: 0.6365. Min Range: -0.6365

Max Act: 0.5145. Min Act: -0.0000

Data Index: 380463 (C4 (Web Text))

Max Activating Token Index: 723

Text #8

Max Range: 0.6365. Min Range: -0.6365

Max Act: 0.5296. Min Act: -0.0001

Data Index: 876300 (C4 (Web Text))

Max Activating Token Index: 894

Text #9

Max Range: 0.6365. Min Range: -0.6365

Max Act: 0.5094. Min Act: -0.0000

Data Index: 1024 (C4 (Web Text))

Max Activating Token Index: 375

Text #10

Max Range: 0.6365. Min Range: -0.6365

Max Act: 0.4977. Min Act: -0.0001

Data Index: 78138 (C4 (Web Text))

Max Activating Token Index: 231

Text #11

Max Range: 0.6365. Min Range: -0.6365

Max Act: 0.4852. Min Act: -0.0001

Data Index: 844871 (C4 (Web Text))

Max Activating Token Index: 936

Text #12

Max Range: 0.6365. Min Range: -0.6365

Max Act: 0.5155. Min Act: -0.0001

Data Index: 1493284 (Python Code)

Max Activating Token Index: 35

Text #13

Max Range: 0.6365. Min Range: -0.6365

Max Act: 0.4858. Min Act: -0.0001

Data Index: 637079 (C4 (Web Text))

Max Activating Token Index: 862

Text #14

Max Range: 0.6365. Min Range: -0.6365

Max Act: 0.4572. Min Act: -0.0000

Data Index: 813461 (C4 (Web Text))

Max Activating Token Index: 837

Text #15

Max Range: 0.6365. Min Range: -0.6365

Max Act: 0.4027. Min Act: -0.0001

Data Index: 1482828 (Python Code)

Max Activating Token Index: 853

Text #16

Max Range: 0.6365. Min Range: -0.6365

Max Act: 0.4416. Min Act: -0.0001

Data Index: 390066 (C4 (Web Text))

Max Activating Token Index: 790

Text #17

Max Range: 0.6365. Min Range: -0.6365

Max Act: 0.4519. Min Act: -0.0001

Data Index: 1613678 (Python Code)

Max Activating Token Index: 468

Text #18

Max Range: 0.6365. Min Range: -0.6365

Max Act: 0.4469. Min Act: -0.0001

Data Index: 1401260 (Python Code)

Max Activating Token Index: 682

Text #19

Max Range: 0.6365. Min Range: -0.6365

Max Act: 0.4436. Min Act: -0.0000

Data Index: 4277 (C4 (Web Text))

Max Activating Token Index: 117