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

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

Neuron 608 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: 0.5380. Min Range: -0.5380

Max Act: 0.5380. Min Act: -0.0000

Data Index: 686790 (C4 (Web Text))

Max Activating Token Index: 898

Text #1

Max Range: 0.5380. Min Range: -0.5380

Max Act: 0.4148. Min Act: -0.0000

Data Index: 377773 (C4 (Web Text))

Max Activating Token Index: 667

Text #2

Max Range: 0.5380. Min Range: -0.5380

Max Act: 0.4303. Min Act: -0.0000

Data Index: 512623 (C4 (Web Text))

Max Activating Token Index: 726

Text #3

Max Range: 0.5380. Min Range: -0.5380

Max Act: 0.5212. Min Act: -0.0000

Data Index: 1094547 (C4 (Web Text))

Max Activating Token Index: 998

Text #4

Max Range: 0.5380. Min Range: -0.5380

Max Act: 0.4331. Min Act: -0.0000

Data Index: 201923 (C4 (Web Text))

Max Activating Token Index: 895

Text #5

Max Range: 0.5380. Min Range: -0.5380

Max Act: 0.3739. Min Act: -0.0000

Data Index: 1263906 (C4 (Web Text))

Max Activating Token Index: 746

Text #6

Max Range: 0.5380. Min Range: -0.5380

Max Act: 0.3577. Min Act: -0.0000

Data Index: 1336318 (C4 (Web Text))

Max Activating Token Index: 347

Text #7

Max Range: 0.5380. Min Range: -0.5380

Max Act: 0.3835. Min Act: -0.0000

Data Index: 177229 (C4 (Web Text))

Max Activating Token Index: 435

Text #8

Max Range: 0.5380. Min Range: -0.5380

Max Act: 0.3108. Min Act: -0.0000

Data Index: 1656230 (Python Code)

Max Activating Token Index: 8

Text #9

Max Range: 0.5380. Min Range: -0.5380

Max Act: 0.3116. Min Act: -0.0000

Data Index: 1418311 (Python Code)

Max Activating Token Index: 9

Text #10

Max Range: 0.5380. Min Range: -0.5380

Max Act: 0.2910. Min Act: -0.0000

Data Index: 674007 (C4 (Web Text))

Max Activating Token Index: 460

Text #11

Max Range: 0.5380. Min Range: -0.5380

Max Act: 0.3004. Min Act: -0.0000

Data Index: 1600574 (Python Code)

Max Activating Token Index: 12

Text #12

Max Range: 0.5380. Min Range: -0.5380

Max Act: 0.2612. Min Act: -0.0000

Data Index: 678109 (C4 (Web Text))

Max Activating Token Index: 135

Text #13

Max Range: 0.5380. Min Range: -0.5380

Max Act: 0.2852. Min Act: -0.0000

Data Index: 1441739 (Python Code)

Max Activating Token Index: 51

Text #14

Max Range: 0.5380. Min Range: -0.5380

Max Act: 0.2840. Min Act: -0.0000

Data Index: 1562752 (Python Code)

Max Activating Token Index: 31

Text #15

Max Range: 0.5380. Min Range: -0.5380

Max Act: 0.2718. Min Act: -0.0000

Data Index: 1649367 (Python Code)

Max Activating Token Index: 12

Text #16

Max Range: 0.5380. Min Range: -0.5380

Max Act: 0.2826. Min Act: -0.0000

Data Index: 78303 (C4 (Web Text))

Max Activating Token Index: 168

Text #17

Max Range: 0.5380. Min Range: -0.5380

Max Act: 0.2634. Min Act: -0.0000

Data Index: 1563877 (Python Code)

Max Activating Token Index: 325

Text #18

Max Range: 0.5380. Min Range: -0.5380

Max Act: 0.2594. Min Act: -0.0000

Data Index: 1467729 (Python Code)

Max Activating Token Index: 22

Text #19

Max Range: 0.5380. Min Range: -0.5380

Max Act: 0.2544. Min Act: -0.0000

Data Index: 1460221 (Python Code)

Max Activating Token Index: 19