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

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

Neuron 72 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: 5.2536. Min Range: -5.2536

Max Act: 5.2536. Min Act: -0.1699

Data Index: 367232 (C4 (Web Text))

Max Activating Token Index: 836

Text #1

Max Range: 5.2536. Min Range: -5.2536

Max Act: 5.1599. Min Act: -0.1700

Data Index: 1195622 (C4 (Web Text))

Max Activating Token Index: 815

Text #2

Max Range: 5.2536. Min Range: -5.2536

Max Act: 5.0192. Min Act: -0.1700

Data Index: 660412 (C4 (Web Text))

Max Activating Token Index: 421

Text #3

Max Range: 5.2536. Min Range: -5.2536

Max Act: 4.9724. Min Act: -0.1700

Data Index: 9575 (C4 (Web Text))

Max Activating Token Index: 572

Text #4

Max Range: 5.2536. Min Range: -5.2536

Max Act: 4.9567. Min Act: -0.1700

Data Index: 138189 (C4 (Web Text))

Max Activating Token Index: 754

Text #5

Max Range: 5.2536. Min Range: -5.2536

Max Act: 4.9255. Min Act: -0.1700

Data Index: 554496 (C4 (Web Text))

Max Activating Token Index: 910

Text #6

Max Range: 5.2536. Min Range: -5.2536

Max Act: 4.8942. Min Act: -0.1700

Data Index: 1266087 (C4 (Web Text))

Max Activating Token Index: 47

Text #7

Max Range: 5.2536. Min Range: -5.2536

Max Act: 4.8474. Min Act: -0.1700

Data Index: 25447 (C4 (Web Text))

Max Activating Token Index: 535

Text #8

Max Range: 5.2536. Min Range: -5.2536

Max Act: 4.8786. Min Act: -0.1700

Data Index: 888184 (C4 (Web Text))

Max Activating Token Index: 524

Text #9

Max Range: 5.2536. Min Range: -5.2536

Max Act: 4.8474. Min Act: -0.1700

Data Index: 952463 (C4 (Web Text))

Max Activating Token Index: 488

Text #10

Max Range: 5.2536. Min Range: -5.2536

Max Act: 4.8786. Min Act: -0.1700

Data Index: 1113270 (C4 (Web Text))

Max Activating Token Index: 994

Text #11

Max Range: 5.2536. Min Range: -5.2536

Max Act: 4.8630. Min Act: -0.1700

Data Index: 1200112 (C4 (Web Text))

Max Activating Token Index: 125

Text #12

Max Range: 5.2536. Min Range: -5.2536

Max Act: 4.8630. Min Act: -0.1700

Data Index: 6932 (C4 (Web Text))

Max Activating Token Index: 50

Text #13

Max Range: 5.2536. Min Range: -5.2536

Max Act: 4.8786. Min Act: -0.1700

Data Index: 1233102 (C4 (Web Text))

Max Activating Token Index: 365

Text #14

Max Range: 5.2536. Min Range: -5.2536

Max Act: 4.7380. Min Act: -0.1700

Data Index: 882010 (C4 (Web Text))

Max Activating Token Index: 219

Text #15

Max Range: 5.2536. Min Range: -5.2536

Max Act: 4.7224. Min Act: -0.1700

Data Index: 153084 (C4 (Web Text))

Max Activating Token Index: 505

Text #16

Max Range: 5.2536. Min Range: -5.2536

Max Act: 4.6130. Min Act: -0.1700

Data Index: 101589 (C4 (Web Text))

Max Activating Token Index: 1002

Text #17

Max Range: 5.2536. Min Range: -5.2536

Max Act: 4.7067. Min Act: -0.1700

Data Index: 1119282 (C4 (Web Text))

Max Activating Token Index: 494

Text #18

Max Range: 5.2536. Min Range: -5.2536

Max Act: 4.7380. Min Act: -0.1700

Data Index: 892490 (C4 (Web Text))

Max Activating Token Index: 135

Text #19

Max Range: 5.2536. Min Range: -5.2536

Max Act: 4.6911. Min Act: -0.1700

Data Index: 1232999 (C4 (Web Text))

Max Activating Token Index: 773