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

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

Neuron 3072 in Layer 0

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0198. Min Act: -0.0003

Data Index: 1285404 (C4 (Web Text))

Max Activating Token Index: 839

Text #1

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0147. Min Act: -0.0003

Data Index: 5084 (C4 (Web Text))

Max Activating Token Index: 165

Text #2

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0146. Min Act: -0.0003

Data Index: 1002566 (C4 (Web Text))

Max Activating Token Index: 548

Text #3

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0135. Min Act: -0.0003

Data Index: 46003 (C4 (Web Text))

Max Activating Token Index: 724

Text #4

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0135. Min Act: -0.0003

Data Index: 1299405 (C4 (Web Text))

Max Activating Token Index: 770

Text #5

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0137. Min Act: -0.0003

Data Index: 681494 (C4 (Web Text))

Max Activating Token Index: 295

Text #6

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0134. Min Act: -0.0003

Data Index: 1346297 (C4 (Web Text))

Max Activating Token Index: 797

Text #7

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0133. Min Act: -0.0002

Data Index: 1406856 (Python Code)

Max Activating Token Index: 18

Text #8

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0132. Min Act: -0.0003

Data Index: 464586 (C4 (Web Text))

Max Activating Token Index: 380

Text #9

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0135. Min Act: -0.0003

Data Index: 1181175 (C4 (Web Text))

Max Activating Token Index: 404

Text #10

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0129. Min Act: -0.0003

Data Index: 1083730 (C4 (Web Text))

Max Activating Token Index: 198

Text #11

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0129. Min Act: -0.0003

Data Index: 1267444 (C4 (Web Text))

Max Activating Token Index: 588

Text #12

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0128. Min Act: -0.0003

Data Index: 978028 (C4 (Web Text))

Max Activating Token Index: 652

Text #13

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0130. Min Act: -0.0003

Data Index: 1033971 (C4 (Web Text))

Max Activating Token Index: 215

Text #14

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0127. Min Act: -0.0003

Data Index: 345011 (C4 (Web Text))

Max Activating Token Index: 761

Text #15

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0126. Min Act: -0.0003

Data Index: 20882 (C4 (Web Text))

Max Activating Token Index: 218

Text #16

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0125. Min Act: -0.0003

Data Index: 1072362 (C4 (Web Text))

Max Activating Token Index: 350

Text #17

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0126. Min Act: -0.0002

Data Index: 1548153 (Python Code)

Max Activating Token Index: 21

Text #18

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0122. Min Act: -0.0003

Data Index: 1341882 (C4 (Web Text))

Max Activating Token Index: 387

Text #19

Max Range: 0.0198. Min Range: -0.0198

Max Act: 0.0121. Min Act: -0.0003

Data Index: 669793 (C4 (Web Text))

Max Activating Token Index: 701