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

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

Neuron 427 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: 5.6989. Min Range: -5.6989

Max Act: 5.6989. Min Act: -0.1700

Data Index: 1236552 (C4 (Web Text))

Max Activating Token Index: 991

Text #1

Max Range: 5.6989. Min Range: -5.6989

Max Act: 5.6364. Min Act: -0.1700

Data Index: 883512 (C4 (Web Text))

Max Activating Token Index: 891

Text #2

Max Range: 5.6989. Min Range: -5.6989

Max Act: 5.5114. Min Act: -0.1700

Data Index: 33321 (C4 (Web Text))

Max Activating Token Index: 44

Text #3

Max Range: 5.6989. Min Range: -5.6989

Max Act: 5.4489. Min Act: -0.1700

Data Index: 394764 (C4 (Web Text))

Max Activating Token Index: 235

Text #4

Max Range: 5.6989. Min Range: -5.6989

Max Act: 5.2926. Min Act: -0.1700

Data Index: 643887 (C4 (Web Text))

Max Activating Token Index: 364

Text #5

Max Range: 5.6989. Min Range: -5.6989

Max Act: 5.2926. Min Act: -0.1700

Data Index: 1025701 (C4 (Web Text))

Max Activating Token Index: 159

Text #6

Max Range: 5.6989. Min Range: -5.6989

Max Act: 5.2926. Min Act: -0.1700

Data Index: 507799 (C4 (Web Text))

Max Activating Token Index: 826

Text #7

Max Range: 5.6989. Min Range: -5.6989

Max Act: 5.3239. Min Act: -0.1700

Data Index: 545436 (C4 (Web Text))

Max Activating Token Index: 917

Text #8

Max Range: 5.6989. Min Range: -5.6989

Max Act: 5.2301. Min Act: -0.1700

Data Index: 261364 (C4 (Web Text))

Max Activating Token Index: 847

Text #9

Max Range: 5.6989. Min Range: -5.6989

Max Act: 5.2301. Min Act: -0.1700

Data Index: 863230 (C4 (Web Text))

Max Activating Token Index: 168

Text #10

Max Range: 5.6989. Min Range: -5.6989

Max Act: 5.2926. Min Act: -0.1700

Data Index: 90555 (C4 (Web Text))

Max Activating Token Index: 903

Text #11

Max Range: 5.6989. Min Range: -5.6989

Max Act: 5.2301. Min Act: -0.1700

Data Index: 433719 (C4 (Web Text))

Max Activating Token Index: 711

Text #12

Max Range: 5.6989. Min Range: -5.6989

Max Act: 5.1989. Min Act: -0.1700

Data Index: 1560619 (Python Code)

Max Activating Token Index: 911

Text #13

Max Range: 5.6989. Min Range: -5.6989

Max Act: 5.2614. Min Act: -0.1700

Data Index: 1499968 (Python Code)

Max Activating Token Index: 49

Text #14

Max Range: 5.6989. Min Range: -5.6989

Max Act: 5.1989. Min Act: -0.1700

Data Index: 71938 (C4 (Web Text))

Max Activating Token Index: 30

Text #15

Max Range: 5.6989. Min Range: -5.6989

Max Act: 5.1989. Min Act: -0.1700

Data Index: 1264823 (C4 (Web Text))

Max Activating Token Index: 498

Text #16

Max Range: 5.6989. Min Range: -5.6989

Max Act: 5.1364. Min Act: -0.1700

Data Index: 1248875 (C4 (Web Text))

Max Activating Token Index: 237

Text #17

Max Range: 5.6989. Min Range: -5.6989

Max Act: 5.1364. Min Act: -0.1700

Data Index: 419992 (C4 (Web Text))

Max Activating Token Index: 358

Text #18

Max Range: 5.6989. Min Range: -5.6989

Max Act: 5.1676. Min Act: -0.1700

Data Index: 259730 (C4 (Web Text))

Max Activating Token Index: 466

Text #19

Max Range: 5.6989. Min Range: -5.6989

Max Act: 5.1364. Min Act: -0.1700

Data Index: 1085200 (C4 (Web Text))

Max Activating Token Index: 405