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

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

Neuron 642 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: 5.4554. Min Range: -5.4554

Max Act: 5.4554. Min Act: -0.1700

Data Index: 925874 (C4 (Web Text))

Max Activating Token Index: 601

Text #1

Max Range: 5.4554. Min Range: -5.4554

Max Act: 5.3929. Min Act: -0.1700

Data Index: 1337915 (C4 (Web Text))

Max Activating Token Index: 987

Text #2

Max Range: 5.4554. Min Range: -5.4554

Max Act: 5.2054. Min Act: -0.1700

Data Index: 1563532 (Python Code)

Max Activating Token Index: 1004

Text #3

Max Range: 5.4554. Min Range: -5.4554

Max Act: 5.2054. Min Act: -0.1700

Data Index: 1634742 (Python Code)

Max Activating Token Index: 897

Text #4

Max Range: 5.4554. Min Range: -5.4554

Max Act: 5.1429. Min Act: -0.1700

Data Index: 1168099 (C4 (Web Text))

Max Activating Token Index: 316

Text #5

Max Range: 5.4554. Min Range: -5.4554

Max Act: 5.1429. Min Act: -0.1700

Data Index: 1621881 (Python Code)

Max Activating Token Index: 629

Text #6

Max Range: 5.4554. Min Range: -5.4554

Max Act: 5.0492. Min Act: -0.1700

Data Index: 575037 (C4 (Web Text))

Max Activating Token Index: 108

Text #7

Max Range: 5.4554. Min Range: -5.4554

Max Act: 5.0804. Min Act: -0.1700

Data Index: 1201346 (C4 (Web Text))

Max Activating Token Index: 305

Text #8

Max Range: 5.4554. Min Range: -5.4554

Max Act: 4.9867. Min Act: -0.1700

Data Index: 473901 (C4 (Web Text))

Max Activating Token Index: 428

Text #9

Max Range: 5.4554. Min Range: -5.4554

Max Act: 5.0492. Min Act: -0.1700

Data Index: 337017 (C4 (Web Text))

Max Activating Token Index: 677

Text #10

Max Range: 5.4554. Min Range: -5.4554

Max Act: 5.0179. Min Act: -0.1700

Data Index: 1327556 (C4 (Web Text))

Max Activating Token Index: 728

Text #11

Max Range: 5.4554. Min Range: -5.4554

Max Act: 5.0492. Min Act: -0.1700

Data Index: 964886 (C4 (Web Text))

Max Activating Token Index: 824

Text #12

Max Range: 5.4554. Min Range: -5.4554

Max Act: 4.9554. Min Act: -0.1700

Data Index: 1492212 (Python Code)

Max Activating Token Index: 80

Text #13

Max Range: 5.4554. Min Range: -5.4554

Max Act: 4.9242. Min Act: -0.1700

Data Index: 1228807 (C4 (Web Text))

Max Activating Token Index: 786

Text #14

Max Range: 5.4554. Min Range: -5.4554

Max Act: 4.9867. Min Act: -0.1700

Data Index: 546532 (C4 (Web Text))

Max Activating Token Index: 804

Text #15

Max Range: 5.4554. Min Range: -5.4554

Max Act: 4.9554. Min Act: -0.1700

Data Index: 228414 (C4 (Web Text))

Max Activating Token Index: 762

Text #16

Max Range: 5.4554. Min Range: -5.4554

Max Act: 4.9867. Min Act: -0.1700

Data Index: 1550971 (Python Code)

Max Activating Token Index: 302

Text #17

Max Range: 5.4554. Min Range: -5.4554

Max Act: 4.9242. Min Act: -0.1700

Data Index: 1272584 (C4 (Web Text))

Max Activating Token Index: 35

Text #18

Max Range: 5.4554. Min Range: -5.4554

Max Act: 4.9554. Min Act: -0.1700

Data Index: 1158146 (C4 (Web Text))

Max Activating Token Index: 455

Text #19

Max Range: 5.4554. Min Range: -5.4554

Max Act: 4.9554. Min Act: -0.1700

Data Index: 216981 (C4 (Web Text))

Max Activating Token Index: 211