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

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

Neuron 839 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.4027. Min Act: -0.1700

Data Index: 697521 (C4 (Web Text))

Max Activating Token Index: 760

Text #1

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.3402. Min Act: -0.1700

Data Index: 861810 (C4 (Web Text))

Max Activating Token Index: 782

Text #2

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.3402. Min Act: -0.1700

Data Index: 5595 (C4 (Web Text))

Max Activating Token Index: 243

Text #3

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.3714. Min Act: -0.1700

Data Index: 629045 (C4 (Web Text))

Max Activating Token Index: 210

Text #4

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.2777. Min Act: -0.1700

Data Index: 1322552 (C4 (Web Text))

Max Activating Token Index: 308

Text #5

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.2152. Min Act: -0.1700

Data Index: 451688 (C4 (Web Text))

Max Activating Token Index: 222

Text #6

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.2464. Min Act: -0.1700

Data Index: 183934 (C4 (Web Text))

Max Activating Token Index: 378

Text #7

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.1839. Min Act: -0.1700

Data Index: 557233 (C4 (Web Text))

Max Activating Token Index: 292

Text #8

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.0277. Min Act: -0.1700

Data Index: 154938 (C4 (Web Text))

Max Activating Token Index: 940

Text #9

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.1527. Min Act: -0.1700

Data Index: 31518 (C4 (Web Text))

Max Activating Token Index: 467

Text #10

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.1527. Min Act: -0.1700

Data Index: 819577 (C4 (Web Text))

Max Activating Token Index: 369

Text #11

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.1839. Min Act: -0.1700

Data Index: 981109 (C4 (Web Text))

Max Activating Token Index: 229

Text #12

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.1214. Min Act: -0.1700

Data Index: 1171542 (C4 (Web Text))

Max Activating Token Index: 243

Text #13

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.0902. Min Act: -0.1700

Data Index: 481761 (C4 (Web Text))

Max Activating Token Index: 327

Text #14

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.0589. Min Act: -0.1700

Data Index: 1098457 (C4 (Web Text))

Max Activating Token Index: 191

Text #15

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.0589. Min Act: -0.1700

Data Index: 95334 (C4 (Web Text))

Max Activating Token Index: 380

Text #16

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.0902. Min Act: -0.1700

Data Index: 934581 (C4 (Web Text))

Max Activating Token Index: 462

Text #17

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.0902. Min Act: -0.1698

Data Index: 228024 (C4 (Web Text))

Max Activating Token Index: 948

Text #18

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.0902. Min Act: -0.1699

Data Index: 799343 (C4 (Web Text))

Max Activating Token Index: 102

Text #19

Max Range: 5.4027. Min Range: -5.4027

Max Act: 5.0589. Min Act: -0.1700

Data Index: 720763 (C4 (Web Text))

Max Activating Token Index: 475