Model: SoLU Model: 10 Layers, 5120 Neurons per Layer

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

Neuron 4537 in Layer 9

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.7451. Min Act: -0.0000

Data Index: 14249 (C4 (Web Text))

Max Activating Token Index: 567

Text #1

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.7907. Min Act: -0.0000

Data Index: 966792 (C4 (Web Text))

Max Activating Token Index: 437

Text #2

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.7513. Min Act: -0.0000

Data Index: 1199491 (C4 (Web Text))

Max Activating Token Index: 83

Text #3

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.7423. Min Act: -0.0000

Data Index: 262665 (C4 (Web Text))

Max Activating Token Index: 748

Text #4

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.7085. Min Act: -0.0000

Data Index: 488939 (C4 (Web Text))

Max Activating Token Index: 197

Text #5

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.7296. Min Act: -0.0000

Data Index: 268781 (C4 (Web Text))

Max Activating Token Index: 141

Text #6

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.8111. Min Act: -0.0000

Data Index: 1138032 (C4 (Web Text))

Max Activating Token Index: 642

Text #7

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.7599. Min Act: -0.0000

Data Index: 847125 (C4 (Web Text))

Max Activating Token Index: 564

Text #8

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.7328. Min Act: -0.0000

Data Index: 239288 (C4 (Web Text))

Max Activating Token Index: 84

Text #9

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.7763. Min Act: -0.0000

Data Index: 501817 (C4 (Web Text))

Max Activating Token Index: 208

Text #10

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.6944. Min Act: -0.0000

Data Index: 690543 (C4 (Web Text))

Max Activating Token Index: 375

Text #11

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.6845. Min Act: -0.0000

Data Index: 640414 (C4 (Web Text))

Max Activating Token Index: 716

Text #12

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.6746. Min Act: -0.0000

Data Index: 820999 (C4 (Web Text))

Max Activating Token Index: 256

Text #13

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.6472. Min Act: -0.0000

Data Index: 238639 (C4 (Web Text))

Max Activating Token Index: 807

Text #14

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.6554. Min Act: -0.0000

Data Index: 1502173 (Python Code)

Max Activating Token Index: 841

Text #15

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.6698. Min Act: -0.0000

Data Index: 1134345 (C4 (Web Text))

Max Activating Token Index: 689

Text #16

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.6486. Min Act: -0.0000

Data Index: 1425364 (Python Code)

Max Activating Token Index: 369

Text #17

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.7687. Min Act: -0.0000

Data Index: 1316675 (C4 (Web Text))

Max Activating Token Index: 618

Text #18

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.7032. Min Act: -0.0000

Data Index: 259481 (C4 (Web Text))

Max Activating Token Index: 66

Text #19

Max Range: 0.8111. Min Range: -0.8111

Max Act: 0.7420. Min Act: -0.0000

Data Index: 878187 (C4 (Web Text))

Max Activating Token Index: 703