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

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

Neuron 1813 in Layer 6

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 5.3002. Min Range: -5.3002

Max Act: 5.3002. Min Act: -0.0000

Data Index: 1572586 (Python Code)

Max Activating Token Index: 667

Text #1

Max Range: 5.3002. Min Range: -5.3002

Max Act: 5.3001. Min Act: -0.0000

Data Index: 1269096 (C4 (Web Text))

Max Activating Token Index: 255

Text #2

Max Range: 5.3002. Min Range: -5.3002

Max Act: 5.1986. Min Act: -0.0000

Data Index: 1442417 (Python Code)

Max Activating Token Index: 457

Text #3

Max Range: 5.3002. Min Range: -5.3002

Max Act: 5.0219. Min Act: -0.0000

Data Index: 1118372 (C4 (Web Text))

Max Activating Token Index: 684

Text #4

Max Range: 5.3002. Min Range: -5.3002

Max Act: 4.8725. Min Act: -0.0001

Data Index: 1390865 (Python Code)

Max Activating Token Index: 964

Text #5

Max Range: 5.3002. Min Range: -5.3002

Max Act: 4.8568. Min Act: -0.0001

Data Index: 943324 (C4 (Web Text))

Max Activating Token Index: 545

Text #6

Max Range: 5.3002. Min Range: -5.3002

Max Act: 5.0362. Min Act: -0.0000

Data Index: 490968 (C4 (Web Text))

Max Activating Token Index: 465

Text #7

Max Range: 5.3002. Min Range: -5.3002

Max Act: 4.8666. Min Act: -0.0000

Data Index: 461197 (C4 (Web Text))

Max Activating Token Index: 920

Text #8

Max Range: 5.3002. Min Range: -5.3002

Max Act: 4.9747. Min Act: -0.0000

Data Index: 1021304 (C4 (Web Text))

Max Activating Token Index: 314

Text #9

Max Range: 5.3002. Min Range: -5.3002

Max Act: 4.9465. Min Act: -0.0000

Data Index: 586659 (C4 (Web Text))

Max Activating Token Index: 902

Text #10

Max Range: 5.3002. Min Range: -5.3002

Max Act: 4.8874. Min Act: -0.0000

Data Index: 4813 (C4 (Web Text))

Max Activating Token Index: 171

Text #11

Max Range: 5.3002. Min Range: -5.3002

Max Act: 4.8726. Min Act: -0.0000

Data Index: 581435 (C4 (Web Text))

Max Activating Token Index: 680

Text #12

Max Range: 5.3002. Min Range: -5.3002

Max Act: 5.0457. Min Act: -0.0000

Data Index: 301185 (C4 (Web Text))

Max Activating Token Index: 395

Text #13

Max Range: 5.3002. Min Range: -5.3002

Max Act: 4.8455. Min Act: -0.0000

Data Index: 1380981 (Python Code)

Max Activating Token Index: 495

Text #14

Max Range: 5.3002. Min Range: -5.3002

Max Act: 4.8477. Min Act: -0.0000

Data Index: 1193116 (C4 (Web Text))

Max Activating Token Index: 1010

Text #15

Max Range: 5.3002. Min Range: -5.3002

Max Act: 4.8100. Min Act: -0.0000

Data Index: 1359927 (Python Code)

Max Activating Token Index: 294

Text #16

Max Range: 5.3002. Min Range: -5.3002

Max Act: 4.6016. Min Act: -0.0000

Data Index: 1280545 (C4 (Web Text))

Max Activating Token Index: 549

Text #17

Max Range: 5.3002. Min Range: -5.3002

Max Act: 4.6006. Min Act: -0.0000

Data Index: 921912 (C4 (Web Text))

Max Activating Token Index: 264

Text #18

Max Range: 5.3002. Min Range: -5.3002

Max Act: 4.7737. Min Act: -0.0000

Data Index: 1333979 (C4 (Web Text))

Max Activating Token Index: 458

Text #19

Max Range: 5.3002. Min Range: -5.3002

Max Act: 4.7805. Min Act: -0.0000

Data Index: 205679 (C4 (Web Text))

Max Activating Token Index: 103