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

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

Neuron 258 in Layer 6

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.5707. Min Act: -0.0000

Data Index: 1025598 (C4 (Web Text))

Max Activating Token Index: 290

Text #1

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.4506. Min Act: -0.0000

Data Index: 267444 (C4 (Web Text))

Max Activating Token Index: 833

Text #2

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.4676. Min Act: -0.0000

Data Index: 531278 (C4 (Web Text))

Max Activating Token Index: 105

Text #3

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.4509. Min Act: -0.0000

Data Index: 106988 (C4 (Web Text))

Max Activating Token Index: 206

Text #4

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.4630. Min Act: -0.0000

Data Index: 793093 (C4 (Web Text))

Max Activating Token Index: 18

Text #5

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.4597. Min Act: -0.0000

Data Index: 915652 (C4 (Web Text))

Max Activating Token Index: 857

Text #6

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.4507. Min Act: -0.0000

Data Index: 1305834 (C4 (Web Text))

Max Activating Token Index: 742

Text #7

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.4484. Min Act: -0.0000

Data Index: 361592 (C4 (Web Text))

Max Activating Token Index: 690

Text #8

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.4398. Min Act: -0.0000

Data Index: 649677 (C4 (Web Text))

Max Activating Token Index: 168

Text #9

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.4331. Min Act: -0.0000

Data Index: 570766 (C4 (Web Text))

Max Activating Token Index: 270

Text #10

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.4383. Min Act: -0.0000

Data Index: 556658 (C4 (Web Text))

Max Activating Token Index: 909

Text #11

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.4166. Min Act: -0.0000

Data Index: 376276 (C4 (Web Text))

Max Activating Token Index: 903

Text #12

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.4331. Min Act: -0.0000

Data Index: 853558 (C4 (Web Text))

Max Activating Token Index: 245

Text #13

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.3998. Min Act: -0.0000

Data Index: 1426265 (Python Code)

Max Activating Token Index: 1001

Text #14

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.4359. Min Act: -0.0000

Data Index: 581892 (C4 (Web Text))

Max Activating Token Index: 236

Text #15

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.3757. Min Act: -0.0000

Data Index: 882850 (C4 (Web Text))

Max Activating Token Index: 816

Text #16

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.4063. Min Act: -0.0000

Data Index: 817808 (C4 (Web Text))

Max Activating Token Index: 709

Text #17

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.3981. Min Act: -0.0000

Data Index: 692688 (C4 (Web Text))

Max Activating Token Index: 67

Text #18

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.4006. Min Act: -0.0000

Data Index: 1142126 (C4 (Web Text))

Max Activating Token Index: 1006

Text #19

Max Range: 0.5707. Min Range: -0.5707

Max Act: 0.4114. Min Act: -0.0000

Data Index: 821017 (C4 (Web Text))

Max Activating Token Index: 553