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

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

Neuron 1583 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: 5.3372. Min Range: -5.3372

Max Act: 5.3372. Min Act: -0.1700

Data Index: 1369145 (Python Code)

Max Activating Token Index: 641

Text #1

Max Range: 5.3372. Min Range: -5.3372

Max Act: 5.3372. Min Act: -0.1700

Data Index: 844506 (C4 (Web Text))

Max Activating Token Index: 829

Text #2

Max Range: 5.3372. Min Range: -5.3372

Max Act: 5.2903. Min Act: -0.1700

Data Index: 1051264 (C4 (Web Text))

Max Activating Token Index: 1010

Text #3

Max Range: 5.3372. Min Range: -5.3372

Max Act: 5.2747. Min Act: -0.1700

Data Index: 1252254 (C4 (Web Text))

Max Activating Token Index: 13

Text #4

Max Range: 5.3372. Min Range: -5.3372

Max Act: 5.2122. Min Act: -0.1700

Data Index: 1170878 (C4 (Web Text))

Max Activating Token Index: 224

Text #5

Max Range: 5.3372. Min Range: -5.3372

Max Act: 5.0872. Min Act: -0.1700

Data Index: 285029 (C4 (Web Text))

Max Activating Token Index: 65

Text #6

Max Range: 5.3372. Min Range: -5.3372

Max Act: 5.1809. Min Act: -0.1700

Data Index: 973598 (C4 (Web Text))

Max Activating Token Index: 144

Text #7

Max Range: 5.3372. Min Range: -5.3372

Max Act: 5.1184. Min Act: -0.1700

Data Index: 1291942 (C4 (Web Text))

Max Activating Token Index: 1003

Text #8

Max Range: 5.3372. Min Range: -5.3372

Max Act: 5.0872. Min Act: -0.1700

Data Index: 472501 (C4 (Web Text))

Max Activating Token Index: 112

Text #9

Max Range: 5.3372. Min Range: -5.3372

Max Act: 5.1028. Min Act: -0.1700

Data Index: 1157659 (C4 (Web Text))

Max Activating Token Index: 150

Text #10

Max Range: 5.3372. Min Range: -5.3372

Max Act: 5.0715. Min Act: -0.1700

Data Index: 547258 (C4 (Web Text))

Max Activating Token Index: 651

Text #11

Max Range: 5.3372. Min Range: -5.3372

Max Act: 5.0559. Min Act: -0.1700

Data Index: 865477 (C4 (Web Text))

Max Activating Token Index: 229

Text #12

Max Range: 5.3372. Min Range: -5.3372

Max Act: 5.0715. Min Act: -0.1700

Data Index: 248925 (C4 (Web Text))

Max Activating Token Index: 396

Text #13

Max Range: 5.3372. Min Range: -5.3372

Max Act: 5.0872. Min Act: -0.1700

Data Index: 851709 (C4 (Web Text))

Max Activating Token Index: 649

Text #14

Max Range: 5.3372. Min Range: -5.3372

Max Act: 5.0090. Min Act: -0.1700

Data Index: 104237 (C4 (Web Text))

Max Activating Token Index: 839

Text #15

Max Range: 5.3372. Min Range: -5.3372

Max Act: 5.0247. Min Act: -0.1700

Data Index: 637027 (C4 (Web Text))

Max Activating Token Index: 260

Text #16

Max Range: 5.3372. Min Range: -5.3372

Max Act: 5.0559. Min Act: -0.1700

Data Index: 861066 (C4 (Web Text))

Max Activating Token Index: 91

Text #17

Max Range: 5.3372. Min Range: -5.3372

Max Act: 5.0559. Min Act: -0.1700

Data Index: 1616438 (Python Code)

Max Activating Token Index: 477

Text #18

Max Range: 5.3372. Min Range: -5.3372

Max Act: 4.9622. Min Act: -0.1700

Data Index: 815794 (C4 (Web Text))

Max Activating Token Index: 984

Text #19

Max Range: 5.3372. Min Range: -5.3372

Max Act: 5.0090. Min Act: -0.1700

Data Index: 1030206 (C4 (Web Text))

Max Activating Token Index: 359