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

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

Neuron 1793 in Layer 0

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 5.3670. Min Range: -5.3670

Max Act: 5.3670. Min Act: -0.1700

Data Index: 283227 (C4 (Web Text))

Max Activating Token Index: 303

Text #1

Max Range: 5.3670. Min Range: -5.3670

Max Act: 5.3357. Min Act: -0.1700

Data Index: 1276417 (C4 (Web Text))

Max Activating Token Index: 931

Text #2

Max Range: 5.3670. Min Range: -5.3670

Max Act: 5.2107. Min Act: -0.1700

Data Index: 1268719 (C4 (Web Text))

Max Activating Token Index: 949

Text #3

Max Range: 5.3670. Min Range: -5.3670

Max Act: 5.1795. Min Act: -0.1700

Data Index: 1515001 (Python Code)

Max Activating Token Index: 76

Text #4

Max Range: 5.3670. Min Range: -5.3670

Max Act: 5.1482. Min Act: -0.1700

Data Index: 1012732 (C4 (Web Text))

Max Activating Token Index: 108

Text #5

Max Range: 5.3670. Min Range: -5.3670

Max Act: 5.0857. Min Act: -0.1700

Data Index: 1272471 (C4 (Web Text))

Max Activating Token Index: 171

Text #6

Max Range: 5.3670. Min Range: -5.3670

Max Act: 5.1170. Min Act: -0.1700

Data Index: 580406 (C4 (Web Text))

Max Activating Token Index: 440

Text #7

Max Range: 5.3670. Min Range: -5.3670

Max Act: 4.9920. Min Act: -0.1700

Data Index: 1148035 (C4 (Web Text))

Max Activating Token Index: 842

Text #8

Max Range: 5.3670. Min Range: -5.3670

Max Act: 5.0232. Min Act: -0.1700

Data Index: 1391953 (Python Code)

Max Activating Token Index: 243

Text #9

Max Range: 5.3670. Min Range: -5.3670

Max Act: 4.9920. Min Act: -0.1700

Data Index: 1576111 (Python Code)

Max Activating Token Index: 12

Text #10

Max Range: 5.3670. Min Range: -5.3670

Max Act: 4.9920. Min Act: -0.1700

Data Index: 1511633 (Python Code)

Max Activating Token Index: 349

Text #11

Max Range: 5.3670. Min Range: -5.3670

Max Act: 4.9920. Min Act: -0.1700

Data Index: 489098 (C4 (Web Text))

Max Activating Token Index: 330

Text #12

Max Range: 5.3670. Min Range: -5.3670

Max Act: 4.9607. Min Act: -0.1700

Data Index: 1206451 (C4 (Web Text))

Max Activating Token Index: 259

Text #13

Max Range: 5.3670. Min Range: -5.3670

Max Act: 4.9920. Min Act: -0.1700

Data Index: 862775 (C4 (Web Text))

Max Activating Token Index: 489

Text #14

Max Range: 5.3670. Min Range: -5.3670

Max Act: 4.9607. Min Act: -0.1700

Data Index: 434922 (C4 (Web Text))

Max Activating Token Index: 374

Text #15

Max Range: 5.3670. Min Range: -5.3670

Max Act: 4.9607. Min Act: -0.1700

Data Index: 1161005 (C4 (Web Text))

Max Activating Token Index: 188

Text #16

Max Range: 5.3670. Min Range: -5.3670

Max Act: 4.9295. Min Act: -0.1700

Data Index: 607441 (C4 (Web Text))

Max Activating Token Index: 900

Text #17

Max Range: 5.3670. Min Range: -5.3670

Max Act: 4.9607. Min Act: -0.1700

Data Index: 224268 (C4 (Web Text))

Max Activating Token Index: 570

Text #18

Max Range: 5.3670. Min Range: -5.3670

Max Act: 4.8982. Min Act: -0.1700

Data Index: 469587 (C4 (Web Text))

Max Activating Token Index: 351

Text #19

Max Range: 5.3670. Min Range: -5.3670

Max Act: 4.8670. Min Act: -0.1700

Data Index: 1214945 (C4 (Web Text))

Max Activating Token Index: 609