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

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

Neuron 440 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.7708. Min Act: -0.0001

Data Index: 297657 (C4 (Web Text))

Max Activating Token Index: 155

Text #1

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.8215. Min Act: -0.0001

Data Index: 408899 (C4 (Web Text))

Max Activating Token Index: 80

Text #2

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.7153. Min Act: -0.0001

Data Index: 1179886 (C4 (Web Text))

Max Activating Token Index: 782

Text #3

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.7575. Min Act: -0.0001

Data Index: 570061 (C4 (Web Text))

Max Activating Token Index: 863

Text #4

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.7509. Min Act: -0.0001

Data Index: 1077418 (C4 (Web Text))

Max Activating Token Index: 84

Text #5

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.7384. Min Act: -0.0001

Data Index: 1216119 (C4 (Web Text))

Max Activating Token Index: 685

Text #6

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.6736. Min Act: -0.0001

Data Index: 1224362 (C4 (Web Text))

Max Activating Token Index: 66

Text #7

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.6112. Min Act: -0.0001

Data Index: 430800 (C4 (Web Text))

Max Activating Token Index: 312

Text #8

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.6458. Min Act: -0.0001

Data Index: 317415 (C4 (Web Text))

Max Activating Token Index: 216

Text #9

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.6423. Min Act: -0.0001

Data Index: 1260760 (C4 (Web Text))

Max Activating Token Index: 895

Text #10

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.6620. Min Act: -0.0001

Data Index: 421215 (C4 (Web Text))

Max Activating Token Index: 755

Text #11

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.6040. Min Act: -0.0001

Data Index: 1209889 (C4 (Web Text))

Max Activating Token Index: 44

Text #12

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.6434. Min Act: -0.0001

Data Index: 608895 (C4 (Web Text))

Max Activating Token Index: 474

Text #13

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.6691. Min Act: -0.0001

Data Index: 699319 (C4 (Web Text))

Max Activating Token Index: 686

Text #14

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.6017. Min Act: -0.0001

Data Index: 1643691 (Python Code)

Max Activating Token Index: 111

Text #15

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.5935. Min Act: -0.0001

Data Index: 39830 (C4 (Web Text))

Max Activating Token Index: 24

Text #16

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.5974. Min Act: -0.0001

Data Index: 763752 (C4 (Web Text))

Max Activating Token Index: 995

Text #17

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.6110. Min Act: -0.0001

Data Index: 789758 (C4 (Web Text))

Max Activating Token Index: 575

Text #18

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.5758. Min Act: -0.0001

Data Index: 987003 (C4 (Web Text))

Max Activating Token Index: 260

Text #19

Max Range: 0.8215. Min Range: -0.8215

Max Act: 0.5427. Min Act: -0.0001

Data Index: 249877 (C4 (Web Text))

Max Activating Token Index: 512