Model: GELU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 1037 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-3l')

Text #0

Max Range: 5.4824. Min Range: -5.4824

Max Act: 5.4824. Min Act: -0.1700

Data Index: 1182971 (C4 (Web Text))

Max Activating Token Index: 459

Text #1

Max Range: 5.4824. Min Range: -5.4824

Max Act: 5.2011. Min Act: -0.1700

Data Index: 980510 (C4 (Web Text))

Max Activating Token Index: 356

Text #2

Max Range: 5.4824. Min Range: -5.4824

Max Act: 5.0136. Min Act: -0.1700

Data Index: 1217275 (C4 (Web Text))

Max Activating Token Index: 900

Text #3

Max Range: 5.4824. Min Range: -5.4824

Max Act: 4.8261. Min Act: -0.1700

Data Index: 636470 (C4 (Web Text))

Max Activating Token Index: 569

Text #4

Max Range: 5.4824. Min Range: -5.4824

Max Act: 4.7949. Min Act: -0.1700

Data Index: 437927 (C4 (Web Text))

Max Activating Token Index: 796

Text #5

Max Range: 5.4824. Min Range: -5.4824

Max Act: 4.7011. Min Act: -0.1700

Data Index: 690156 (C4 (Web Text))

Max Activating Token Index: 566

Text #6

Max Range: 5.4824. Min Range: -5.4824

Max Act: 4.6699. Min Act: -0.1700

Data Index: 531860 (C4 (Web Text))

Max Activating Token Index: 473

Text #7

Max Range: 5.4824. Min Range: -5.4824

Max Act: 4.6386. Min Act: -0.1700

Data Index: 727953 (C4 (Web Text))

Max Activating Token Index: 213

Text #8

Max Range: 5.4824. Min Range: -5.4824

Max Act: 4.6699. Min Act: -0.1700

Data Index: 483140 (C4 (Web Text))

Max Activating Token Index: 200

Text #9

Max Range: 5.4824. Min Range: -5.4824

Max Act: 4.6699. Min Act: -0.1700

Data Index: 975823 (C4 (Web Text))

Max Activating Token Index: 931

Text #10

Max Range: 5.4824. Min Range: -5.4824

Max Act: 4.6386. Min Act: -0.1700

Data Index: 611503 (C4 (Web Text))

Max Activating Token Index: 546

Text #11

Max Range: 5.4824. Min Range: -5.4824

Max Act: 4.7011. Min Act: -0.1700

Data Index: 254331 (C4 (Web Text))

Max Activating Token Index: 1018

Text #12

Max Range: 5.4824. Min Range: -5.4824

Max Act: 4.6074. Min Act: -0.1700

Data Index: 1130313 (C4 (Web Text))

Max Activating Token Index: 988

Text #13

Max Range: 5.4824. Min Range: -5.4824

Max Act: 4.6699. Min Act: -0.1700

Data Index: 468408 (C4 (Web Text))

Max Activating Token Index: 665

Text #14

Max Range: 5.4824. Min Range: -5.4824

Max Act: 4.5761. Min Act: -0.1700

Data Index: 874506 (C4 (Web Text))

Max Activating Token Index: 976

Text #15

Max Range: 5.4824. Min Range: -5.4824

Max Act: 4.6074. Min Act: -0.1700

Data Index: 617124 (C4 (Web Text))

Max Activating Token Index: 325

Text #16

Max Range: 5.4824. Min Range: -5.4824

Max Act: 4.6699. Min Act: -0.1700

Data Index: 1293812 (C4 (Web Text))

Max Activating Token Index: 916

Text #17

Max Range: 5.4824. Min Range: -5.4824

Max Act: 4.5761. Min Act: -0.1700

Data Index: 200233 (C4 (Web Text))

Max Activating Token Index: 144

Text #18

Max Range: 5.4824. Min Range: -5.4824

Max Act: 4.6074. Min Act: -0.1700

Data Index: 927852 (C4 (Web Text))

Max Activating Token Index: 991

Text #19

Max Range: 5.4824. Min Range: -5.4824

Max Act: 4.6074. Min Act: -0.1700

Data Index: 379516 (C4 (Web Text))

Max Activating Token Index: 238