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

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

Neuron 1978 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.6555. Min Act: -0.1700

Data Index: 794867 (C4 (Web Text))

Max Activating Token Index: 138

Text #1

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.5305. Min Act: -0.1700

Data Index: 779266 (C4 (Web Text))

Max Activating Token Index: 775

Text #2

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.5305. Min Act: -0.1700

Data Index: 963797 (C4 (Web Text))

Max Activating Token Index: 112

Text #3

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.5618. Min Act: -0.1700

Data Index: 1158293 (C4 (Web Text))

Max Activating Token Index: 714

Text #4

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.4993. Min Act: -0.1700

Data Index: 1179564 (C4 (Web Text))

Max Activating Token Index: 934

Text #5

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.4993. Min Act: -0.1700

Data Index: 1213536 (C4 (Web Text))

Max Activating Token Index: 478

Text #6

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.4368. Min Act: -0.1700

Data Index: 1297810 (C4 (Web Text))

Max Activating Token Index: 579

Text #7

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.4993. Min Act: -0.1700

Data Index: 446882 (C4 (Web Text))

Max Activating Token Index: 738

Text #8

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.4680. Min Act: -0.1700

Data Index: 107918 (C4 (Web Text))

Max Activating Token Index: 917

Text #9

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.4680. Min Act: -0.1700

Data Index: 826452 (C4 (Web Text))

Max Activating Token Index: 613

Text #10

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.4680. Min Act: -0.1700

Data Index: 851223 (C4 (Web Text))

Max Activating Token Index: 803

Text #11

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.4993. Min Act: -0.1700

Data Index: 968800 (C4 (Web Text))

Max Activating Token Index: 580

Text #12

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.4368. Min Act: -0.1700

Data Index: 1071468 (C4 (Web Text))

Max Activating Token Index: 887

Text #13

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.4680. Min Act: -0.1700

Data Index: 1052363 (C4 (Web Text))

Max Activating Token Index: 71

Text #14

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.4368. Min Act: -0.1700

Data Index: 23000 (C4 (Web Text))

Max Activating Token Index: 586

Text #15

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.4680. Min Act: -0.1700

Data Index: 1587852 (Python Code)

Max Activating Token Index: 331

Text #16

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.4368. Min Act: -0.1700

Data Index: 337789 (C4 (Web Text))

Max Activating Token Index: 771

Text #17

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.4368. Min Act: -0.1700

Data Index: 1156884 (C4 (Web Text))

Max Activating Token Index: 433

Text #18

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.4055. Min Act: -0.1700

Data Index: 1175468 (C4 (Web Text))

Max Activating Token Index: 909

Text #19

Max Range: 5.6555. Min Range: -5.6555

Max Act: 5.3743. Min Act: -0.1700

Data Index: 1601272 (Python Code)

Max Activating Token Index: 23