Model: SoLU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 601 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-3l')

Text #0

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0221. Min Act: -0.0006

Data Index: 671844 (C4 (Web Text))

Max Activating Token Index: 217

Text #1

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0218. Min Act: -0.0006

Data Index: 1008180 (C4 (Web Text))

Max Activating Token Index: 161

Text #2

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0209. Min Act: -0.0006

Data Index: 283434 (C4 (Web Text))

Max Activating Token Index: 798

Text #3

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0202. Min Act: -0.0006

Data Index: 709626 (C4 (Web Text))

Max Activating Token Index: 274

Text #4

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0197. Min Act: -0.0006

Data Index: 1002966 (C4 (Web Text))

Max Activating Token Index: 112

Text #5

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0200. Min Act: -0.0006

Data Index: 822434 (C4 (Web Text))

Max Activating Token Index: 585

Text #6

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0189. Min Act: -0.0006

Data Index: 914511 (C4 (Web Text))

Max Activating Token Index: 722

Text #7

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0179. Min Act: -0.0006

Data Index: 674510 (C4 (Web Text))

Max Activating Token Index: 797

Text #8

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0191. Min Act: -0.0006

Data Index: 506829 (C4 (Web Text))

Max Activating Token Index: 831

Text #9

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0204. Min Act: -0.0006

Data Index: 1615777 (Python Code)

Max Activating Token Index: 640

Text #10

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0190. Min Act: -0.0006

Data Index: 705950 (C4 (Web Text))

Max Activating Token Index: 724

Text #11

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0189. Min Act: -0.0006

Data Index: 1420679 (Python Code)

Max Activating Token Index: 948

Text #12

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0185. Min Act: -0.0006

Data Index: 824516 (C4 (Web Text))

Max Activating Token Index: 719

Text #13

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0172. Min Act: -0.0006

Data Index: 1656040 (Python Code)

Max Activating Token Index: 492

Text #14

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0182. Min Act: -0.0006

Data Index: 1186154 (C4 (Web Text))

Max Activating Token Index: 209

Text #15

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0190. Min Act: -0.0006

Data Index: 1154384 (C4 (Web Text))

Max Activating Token Index: 625

Text #16

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0174. Min Act: -0.0006

Data Index: 1038380 (C4 (Web Text))

Max Activating Token Index: 576

Text #17

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0182. Min Act: -0.0006

Data Index: 1482852 (Python Code)

Max Activating Token Index: 167

Text #18

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0189. Min Act: -0.0006

Data Index: 1040404 (C4 (Web Text))

Max Activating Token Index: 383

Text #19

Max Range: 0.0221. Min Range: -0.0221

Max Act: 0.0180. Min Act: -0.0006

Data Index: 1593916 (Python Code)

Max Activating Token Index: 665