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

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

Neuron 2037 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: 8.4425. Min Range: -8.4425

Max Act: 8.4425. Min Act: -0.1700

Data Index: 1250548 (C4 (Web Text))

Max Activating Token Index: 706

Text #1

Max Range: 8.4425. Min Range: -8.4425

Max Act: 8.4113. Min Act: -0.1700

Data Index: 1581042 (Python Code)

Max Activating Token Index: 724

Text #2

Max Range: 8.4425. Min Range: -8.4425

Max Act: 8.3175. Min Act: -0.1700

Data Index: 1425638 (Python Code)

Max Activating Token Index: 801

Text #3

Max Range: 8.4425. Min Range: -8.4425

Max Act: 8.1925. Min Act: -0.1700

Data Index: 1476647 (Python Code)

Max Activating Token Index: 955

Text #4

Max Range: 8.4425. Min Range: -8.4425

Max Act: 8.0988. Min Act: -0.1700

Data Index: 1398907 (Python Code)

Max Activating Token Index: 914

Text #5

Max Range: 8.4425. Min Range: -8.4425

Max Act: 7.8800. Min Act: -0.1700

Data Index: 1389107 (Python Code)

Max Activating Token Index: 814

Text #6

Max Range: 8.4425. Min Range: -8.4425

Max Act: 7.8175. Min Act: -0.1700

Data Index: 231288 (C4 (Web Text))

Max Activating Token Index: 932

Text #7

Max Range: 8.4425. Min Range: -8.4425

Max Act: 7.8488. Min Act: -0.1700

Data Index: 1387669 (Python Code)

Max Activating Token Index: 737

Text #8

Max Range: 8.4425. Min Range: -8.4425

Max Act: 7.8175. Min Act: -0.1700

Data Index: 1498386 (Python Code)

Max Activating Token Index: 846

Text #9

Max Range: 8.4425. Min Range: -8.4425

Max Act: 8.0050. Min Act: -0.1700

Data Index: 1459492 (Python Code)

Max Activating Token Index: 467

Text #10

Max Range: 8.4425. Min Range: -8.4425

Max Act: 7.8488. Min Act: -0.1700

Data Index: 1067058 (C4 (Web Text))

Max Activating Token Index: 296

Text #11

Max Range: 8.4425. Min Range: -8.4425

Max Act: 7.8175. Min Act: -0.1700

Data Index: 399447 (C4 (Web Text))

Max Activating Token Index: 830

Text #12

Max Range: 8.4425. Min Range: -8.4425

Max Act: 7.7863. Min Act: -0.1700

Data Index: 1364316 (Python Code)

Max Activating Token Index: 797

Text #13

Max Range: 8.4425. Min Range: -8.4425

Max Act: 7.8175. Min Act: -0.1700

Data Index: 722617 (C4 (Web Text))

Max Activating Token Index: 668

Text #14

Max Range: 8.4425. Min Range: -8.4425

Max Act: 7.8488. Min Act: -0.1700

Data Index: 1601364 (Python Code)

Max Activating Token Index: 568

Text #15

Max Range: 8.4425. Min Range: -8.4425

Max Act: 7.7550. Min Act: -0.1700

Data Index: 1652816 (Python Code)

Max Activating Token Index: 940

Text #16

Max Range: 8.4425. Min Range: -8.4425

Max Act: 7.7550. Min Act: -0.1700

Data Index: 1625588 (Python Code)

Max Activating Token Index: 922

Text #17

Max Range: 8.4425. Min Range: -8.4425

Max Act: 7.7863. Min Act: -0.1700

Data Index: 1593885 (Python Code)

Max Activating Token Index: 480

Text #18

Max Range: 8.4425. Min Range: -8.4425

Max Act: 7.7863. Min Act: -0.1700

Data Index: 1445488 (Python Code)

Max Activating Token Index: 559

Text #19

Max Range: 8.4425. Min Range: -8.4425

Max Act: 7.6925. Min Act: -0.1700

Data Index: 1573774 (Python Code)

Max Activating Token Index: 412