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

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

Neuron 854 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: 3.7538. Min Range: -3.7538

Max Act: 3.7538. Min Act: -0.1700

Data Index: 869535 (C4 (Web Text))

Max Activating Token Index: 787

Text #1

Max Range: 3.7538. Min Range: -3.7538

Max Act: 3.7225. Min Act: -0.1700

Data Index: 1089754 (C4 (Web Text))

Max Activating Token Index: 746

Text #2

Max Range: 3.7538. Min Range: -3.7538

Max Act: 3.6912. Min Act: -0.1700

Data Index: 228935 (C4 (Web Text))

Max Activating Token Index: 312

Text #3

Max Range: 3.7538. Min Range: -3.7538

Max Act: 3.6912. Min Act: -0.1700

Data Index: 924538 (C4 (Web Text))

Max Activating Token Index: 400

Text #4

Max Range: 3.7538. Min Range: -3.7538

Max Act: 3.6599. Min Act: -0.1700

Data Index: 142657 (C4 (Web Text))

Max Activating Token Index: 700

Text #5

Max Range: 3.7538. Min Range: -3.7538

Max Act: 3.6286. Min Act: -0.1700

Data Index: 1114268 (C4 (Web Text))

Max Activating Token Index: 875

Text #6

Max Range: 3.7538. Min Range: -3.7538

Max Act: 3.6599. Min Act: -0.1700

Data Index: 1251651 (C4 (Web Text))

Max Activating Token Index: 728

Text #7

Max Range: 3.7538. Min Range: -3.7538

Max Act: 3.5973. Min Act: -0.1700

Data Index: 486965 (C4 (Web Text))

Max Activating Token Index: 853

Text #8

Max Range: 3.7538. Min Range: -3.7538

Max Act: 3.5659. Min Act: -0.1700

Data Index: 1086455 (C4 (Web Text))

Max Activating Token Index: 717

Text #9

Max Range: 3.7538. Min Range: -3.7538

Max Act: 3.5659. Min Act: -0.1700

Data Index: 1511205 (Python Code)

Max Activating Token Index: 649

Text #10

Max Range: 3.7538. Min Range: -3.7538

Max Act: 3.5659. Min Act: -0.1700

Data Index: 757395 (C4 (Web Text))

Max Activating Token Index: 958

Text #11

Max Range: 3.7538. Min Range: -3.7538

Max Act: 3.5346. Min Act: -0.1700

Data Index: 1201029 (C4 (Web Text))

Max Activating Token Index: 899

Text #12

Max Range: 3.7538. Min Range: -3.7538

Max Act: 3.5033. Min Act: -0.1700

Data Index: 470330 (C4 (Web Text))

Max Activating Token Index: 658

Text #13

Max Range: 3.7538. Min Range: -3.7538

Max Act: 3.5346. Min Act: -0.1700

Data Index: 229743 (C4 (Web Text))

Max Activating Token Index: 318

Text #14

Max Range: 3.7538. Min Range: -3.7538

Max Act: 3.5346. Min Act: -0.1700

Data Index: 1629160 (Python Code)

Max Activating Token Index: 543

Text #15

Max Range: 3.7538. Min Range: -3.7538

Max Act: 3.4719. Min Act: -0.1700

Data Index: 116764 (C4 (Web Text))

Max Activating Token Index: 837

Text #16

Max Range: 3.7538. Min Range: -3.7538

Max Act: 3.4406. Min Act: -0.1700

Data Index: 1477954 (Python Code)

Max Activating Token Index: 742

Text #17

Max Range: 3.7538. Min Range: -3.7538

Max Act: 3.4092. Min Act: -0.1700

Data Index: 1370094 (Python Code)

Max Activating Token Index: 450

Text #18

Max Range: 3.7538. Min Range: -3.7538

Max Act: 3.4092. Min Act: -0.1700

Data Index: 1524846 (Python Code)

Max Activating Token Index: 783

Text #19

Max Range: 3.7538. Min Range: -3.7538

Max Act: 3.4406. Min Act: -0.1700

Data Index: 1457371 (Python Code)

Max Activating Token Index: 566