Model: GELU Model: 4 Layers, 2048 Neurons per Layer

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

Neuron 678 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-4l')

Text #0

Max Range: 5.4165. Min Range: -5.4165

Max Act: 5.3540. Min Act: -0.1700

Data Index: 835967 (C4 (Web Text))

Max Activating Token Index: 998

Text #1

Max Range: 5.4165. Min Range: -5.4165

Max Act: 5.4165. Min Act: -0.1700

Data Index: 549237 (C4 (Web Text))

Max Activating Token Index: 134

Text #2

Max Range: 5.4165. Min Range: -5.4165

Max Act: 5.1665. Min Act: -0.1700

Data Index: 1375953 (Python Code)

Max Activating Token Index: 477

Text #3

Max Range: 5.4165. Min Range: -5.4165

Max Act: 5.0806. Min Act: -0.1699

Data Index: 1239492 (C4 (Web Text))

Max Activating Token Index: 859

Text #4

Max Range: 5.4165. Min Range: -5.4165

Max Act: 5.1353. Min Act: -0.1699

Data Index: 649077 (C4 (Web Text))

Max Activating Token Index: 246

Text #5

Max Range: 5.4165. Min Range: -5.4165

Max Act: 5.0415. Min Act: -0.1700

Data Index: 1627540 (Python Code)

Max Activating Token Index: 849

Text #6

Max Range: 5.4165. Min Range: -5.4165

Max Act: 5.0650. Min Act: -0.1700

Data Index: 34091 (C4 (Web Text))

Max Activating Token Index: 69

Text #7

Max Range: 5.4165. Min Range: -5.4165

Max Act: 5.0259. Min Act: -0.1700

Data Index: 219322 (C4 (Web Text))

Max Activating Token Index: 40

Text #8

Max Range: 5.4165. Min Range: -5.4165

Max Act: 5.0415. Min Act: -0.1700

Data Index: 315912 (C4 (Web Text))

Max Activating Token Index: 865

Text #9

Max Range: 5.4165. Min Range: -5.4165

Max Act: 4.9946. Min Act: -0.1700

Data Index: 808445 (C4 (Web Text))

Max Activating Token Index: 462

Text #10

Max Range: 5.4165. Min Range: -5.4165

Max Act: 5.0025. Min Act: -0.1700

Data Index: 702295 (C4 (Web Text))

Max Activating Token Index: 188

Text #11

Max Range: 5.4165. Min Range: -5.4165

Max Act: 4.9790. Min Act: -0.1700

Data Index: 837846 (C4 (Web Text))

Max Activating Token Index: 42

Text #12

Max Range: 5.4165. Min Range: -5.4165

Max Act: 4.9556. Min Act: -0.1700

Data Index: 297533 (C4 (Web Text))

Max Activating Token Index: 752

Text #13

Max Range: 5.4165. Min Range: -5.4165

Max Act: 4.9321. Min Act: -0.1700

Data Index: 386338 (C4 (Web Text))

Max Activating Token Index: 48

Text #14

Max Range: 5.4165. Min Range: -5.4165

Max Act: 4.9478. Min Act: -0.1700

Data Index: 26087 (C4 (Web Text))

Max Activating Token Index: 597

Text #15

Max Range: 5.4165. Min Range: -5.4165

Max Act: 4.9009. Min Act: -0.1699

Data Index: 1262431 (C4 (Web Text))

Max Activating Token Index: 482

Text #16

Max Range: 5.4165. Min Range: -5.4165

Max Act: 4.9009. Min Act: -0.1700

Data Index: 29143 (C4 (Web Text))

Max Activating Token Index: 396

Text #17

Max Range: 5.4165. Min Range: -5.4165

Max Act: 4.9165. Min Act: -0.1699

Data Index: 677789 (C4 (Web Text))

Max Activating Token Index: 43

Text #18

Max Range: 5.4165. Min Range: -5.4165

Max Act: 4.9165. Min Act: -0.1699

Data Index: 898277 (C4 (Web Text))

Max Activating Token Index: 422

Text #19

Max Range: 5.4165. Min Range: -5.4165

Max Act: 4.9009. Min Act: -0.1700

Data Index: 209080 (C4 (Web Text))

Max Activating Token Index: 73