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

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

Neuron 1050 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: 3.6966. Min Range: -3.6966

Max Act: 3.6966. Min Act: -0.1700

Data Index: 36822 (C4 (Web Text))

Max Activating Token Index: 988

Text #1

Max Range: 3.6966. Min Range: -3.6966

Max Act: 3.6966. Min Act: -0.1699

Data Index: 1211673 (C4 (Web Text))

Max Activating Token Index: 971

Text #2

Max Range: 3.6966. Min Range: -3.6966

Max Act: 3.6966. Min Act: -0.1700

Data Index: 960740 (C4 (Web Text))

Max Activating Token Index: 370

Text #3

Max Range: 3.6966. Min Range: -3.6966

Max Act: 3.6027. Min Act: -0.1700

Data Index: 464168 (C4 (Web Text))

Max Activating Token Index: 919

Text #4

Max Range: 3.6966. Min Range: -3.6966

Max Act: 3.6027. Min Act: -0.1700

Data Index: 1496659 (Python Code)

Max Activating Token Index: 173

Text #5

Max Range: 3.6966. Min Range: -3.6966

Max Act: 3.6183. Min Act: -0.1700

Data Index: 949177 (C4 (Web Text))

Max Activating Token Index: 920

Text #6

Max Range: 3.6966. Min Range: -3.6966

Max Act: 3.6027. Min Act: -0.1699

Data Index: 1232509 (C4 (Web Text))

Max Activating Token Index: 910

Text #7

Max Range: 3.6966. Min Range: -3.6966

Max Act: 3.5557. Min Act: -0.1700

Data Index: 1162182 (C4 (Web Text))

Max Activating Token Index: 752

Text #8

Max Range: 3.6966. Min Range: -3.6966

Max Act: 3.5244. Min Act: -0.1700

Data Index: 1219905 (C4 (Web Text))

Max Activating Token Index: 719

Text #9

Max Range: 3.6966. Min Range: -3.6966

Max Act: 3.5557. Min Act: -0.1700

Data Index: 1075054 (C4 (Web Text))

Max Activating Token Index: 788

Text #10

Max Range: 3.6966. Min Range: -3.6966

Max Act: 3.5400. Min Act: -0.1700

Data Index: 216704 (C4 (Web Text))

Max Activating Token Index: 325

Text #11

Max Range: 3.6966. Min Range: -3.6966

Max Act: 3.5870. Min Act: -0.1700

Data Index: 1004793 (C4 (Web Text))

Max Activating Token Index: 901

Text #12

Max Range: 3.6966. Min Range: -3.6966

Max Act: 3.5087. Min Act: -0.1700

Data Index: 1079126 (C4 (Web Text))

Max Activating Token Index: 849

Text #13

Max Range: 3.6966. Min Range: -3.6966

Max Act: 3.5244. Min Act: -0.1700

Data Index: 320499 (C4 (Web Text))

Max Activating Token Index: 813

Text #14

Max Range: 3.6966. Min Range: -3.6966

Max Act: 3.5244. Min Act: -0.1700

Data Index: 1062361 (C4 (Web Text))

Max Activating Token Index: 350

Text #15

Max Range: 3.6966. Min Range: -3.6966

Max Act: 3.5244. Min Act: -0.1700

Data Index: 574768 (C4 (Web Text))

Max Activating Token Index: 984

Text #16

Max Range: 3.6966. Min Range: -3.6966

Max Act: 3.4930. Min Act: -0.1700

Data Index: 944147 (C4 (Web Text))

Max Activating Token Index: 508

Text #17

Max Range: 3.6966. Min Range: -3.6966

Max Act: 3.4930. Min Act: -0.1700

Data Index: 29339 (C4 (Web Text))

Max Activating Token Index: 547

Text #18

Max Range: 3.6966. Min Range: -3.6966

Max Act: 3.5244. Min Act: -0.1700

Data Index: 353192 (C4 (Web Text))

Max Activating Token Index: 978

Text #19

Max Range: 3.6966. Min Range: -3.6966

Max Act: 3.4930. Min Act: -0.1700

Data Index: 1212635 (C4 (Web Text))

Max Activating Token Index: 682