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

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

Neuron 199 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 6.0187. Min Range: -6.0187

Max Act: 6.0187. Min Act: -0.0000

Data Index: 385579 (C4 (Web Text))

Max Activating Token Index: 850

Text #1

Max Range: 6.0187. Min Range: -6.0187

Max Act: 4.6466. Min Act: -0.0000

Data Index: 714927 (C4 (Web Text))

Max Activating Token Index: 338

Text #2

Max Range: 6.0187. Min Range: -6.0187

Max Act: 4.3143. Min Act: -0.0000

Data Index: 290099 (C4 (Web Text))

Max Activating Token Index: 731

Text #3

Max Range: 6.0187. Min Range: -6.0187

Max Act: 3.3727. Min Act: -0.0000

Data Index: 423529 (C4 (Web Text))

Max Activating Token Index: 116

Text #4

Max Range: 6.0187. Min Range: -6.0187

Max Act: 3.3723. Min Act: -0.0000

Data Index: 1319030 (C4 (Web Text))

Max Activating Token Index: 328

Text #5

Max Range: 6.0187. Min Range: -6.0187

Max Act: 3.3075. Min Act: -0.0000

Data Index: 433567 (C4 (Web Text))

Max Activating Token Index: 280

Text #6

Max Range: 6.0187. Min Range: -6.0187

Max Act: 3.2514. Min Act: -0.0000

Data Index: 1161689 (C4 (Web Text))

Max Activating Token Index: 150

Text #7

Max Range: 6.0187. Min Range: -6.0187

Max Act: 3.0851. Min Act: -0.0000

Data Index: 1001053 (C4 (Web Text))

Max Activating Token Index: 545

Text #8

Max Range: 6.0187. Min Range: -6.0187

Max Act: 2.7687. Min Act: -0.0000

Data Index: 741445 (C4 (Web Text))

Max Activating Token Index: 989

Text #9

Max Range: 6.0187. Min Range: -6.0187

Max Act: 3.5862. Min Act: -0.0000

Data Index: 411075 (C4 (Web Text))

Max Activating Token Index: 590

Text #10

Max Range: 6.0187. Min Range: -6.0187

Max Act: 3.0128. Min Act: -0.0000

Data Index: 41201 (C4 (Web Text))

Max Activating Token Index: 271

Text #11

Max Range: 6.0187. Min Range: -6.0187

Max Act: 2.9795. Min Act: -0.0000

Data Index: 917352 (C4 (Web Text))

Max Activating Token Index: 655

Text #12

Max Range: 6.0187. Min Range: -6.0187

Max Act: 2.5362. Min Act: -0.0000

Data Index: 989973 (C4 (Web Text))

Max Activating Token Index: 214

Text #13

Max Range: 6.0187. Min Range: -6.0187

Max Act: 3.0543. Min Act: -0.0000

Data Index: 467100 (C4 (Web Text))

Max Activating Token Index: 167

Text #14

Max Range: 6.0187. Min Range: -6.0187

Max Act: 2.7435. Min Act: -0.0000

Data Index: 892664 (C4 (Web Text))

Max Activating Token Index: 364

Text #15

Max Range: 6.0187. Min Range: -6.0187

Max Act: 2.7556. Min Act: -0.0000

Data Index: 492234 (C4 (Web Text))

Max Activating Token Index: 731

Text #16

Max Range: 6.0187. Min Range: -6.0187

Max Act: 3.3111. Min Act: -0.0000

Data Index: 564080 (C4 (Web Text))

Max Activating Token Index: 726

Text #17

Max Range: 6.0187. Min Range: -6.0187

Max Act: 2.7119. Min Act: -0.0000

Data Index: 505372 (C4 (Web Text))

Max Activating Token Index: 636

Text #18

Max Range: 6.0187. Min Range: -6.0187

Max Act: 2.9953. Min Act: -0.0000

Data Index: 229216 (C4 (Web Text))

Max Activating Token Index: 512

Text #19

Max Range: 6.0187. Min Range: -6.0187

Max Act: 2.4193. Min Act: -0.0000

Data Index: 999027 (C4 (Web Text))

Max Activating Token Index: 262