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

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

Neuron 1973 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.5883. Min Act: -0.1700

Data Index: 845217 (C4 (Web Text))

Max Activating Token Index: 397

Text #1

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.4630. Min Act: -0.1700

Data Index: 595673 (C4 (Web Text))

Max Activating Token Index: 389

Text #2

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.4630. Min Act: -0.1700

Data Index: 835815 (C4 (Web Text))

Max Activating Token Index: 1012

Text #3

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.4316. Min Act: -0.1700

Data Index: 1604190 (Python Code)

Max Activating Token Index: 843

Text #4

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.4003. Min Act: -0.1700

Data Index: 388368 (C4 (Web Text))

Max Activating Token Index: 582

Text #5

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.4316. Min Act: -0.1699

Data Index: 1283752 (C4 (Web Text))

Max Activating Token Index: 856

Text #6

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.4003. Min Act: -0.1700

Data Index: 1644264 (Python Code)

Max Activating Token Index: 122

Text #7

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.3689. Min Act: -0.1700

Data Index: 57974 (C4 (Web Text))

Max Activating Token Index: 25

Text #8

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.3689. Min Act: -0.1700

Data Index: 1612098 (Python Code)

Max Activating Token Index: 132

Text #9

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.3689. Min Act: -0.1700

Data Index: 1633300 (Python Code)

Max Activating Token Index: 227

Text #10

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.3689. Min Act: -0.1699

Data Index: 1525561 (Python Code)

Max Activating Token Index: 613

Text #11

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.3689. Min Act: -0.1700

Data Index: 1588967 (Python Code)

Max Activating Token Index: 91

Text #12

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.3375. Min Act: -0.1700

Data Index: 674744 (C4 (Web Text))

Max Activating Token Index: 306

Text #13

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.3689. Min Act: -0.1700

Data Index: 590983 (C4 (Web Text))

Max Activating Token Index: 780

Text #14

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.3689. Min Act: -0.1700

Data Index: 460385 (C4 (Web Text))

Max Activating Token Index: 366

Text #15

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.3689. Min Act: -0.1700

Data Index: 656931 (C4 (Web Text))

Max Activating Token Index: 962

Text #16

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.3689. Min Act: -0.1700

Data Index: 1587890 (Python Code)

Max Activating Token Index: 527

Text #17

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.3375. Min Act: -0.1700

Data Index: 1640198 (Python Code)

Max Activating Token Index: 419

Text #18

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.3375. Min Act: -0.1700

Data Index: 1603910 (Python Code)

Max Activating Token Index: 940

Text #19

Max Range: 3.5883. Min Range: -3.5883

Max Act: 3.3689. Min Act: -0.1700

Data Index: 1601955 (Python Code)

Max Activating Token Index: 939