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

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

Neuron 576 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.8550. Min Act: -0.1700

Data Index: 439949 (C4 (Web Text))

Max Activating Token Index: 868

Text #1

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.8237. Min Act: -0.1700

Data Index: 1523744 (Python Code)

Max Activating Token Index: 794

Text #2

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.7924. Min Act: -0.1700

Data Index: 1399661 (Python Code)

Max Activating Token Index: 955

Text #3

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.7298. Min Act: -0.1700

Data Index: 1592610 (Python Code)

Max Activating Token Index: 444

Text #4

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.7298. Min Act: -0.1700

Data Index: 1548863 (Python Code)

Max Activating Token Index: 485

Text #5

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.7611. Min Act: -0.1700

Data Index: 1609984 (Python Code)

Max Activating Token Index: 376

Text #6

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.7298. Min Act: -0.1700

Data Index: 39732 (C4 (Web Text))

Max Activating Token Index: 329

Text #7

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.7298. Min Act: -0.1700

Data Index: 1522050 (Python Code)

Max Activating Token Index: 256

Text #8

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.6986. Min Act: -0.1700

Data Index: 1407089 (Python Code)

Max Activating Token Index: 972

Text #9

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.6986. Min Act: -0.1700

Data Index: 1511567 (Python Code)

Max Activating Token Index: 352

Text #10

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.6673. Min Act: -0.1700

Data Index: 1511308 (Python Code)

Max Activating Token Index: 708

Text #11

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.6986. Min Act: -0.1700

Data Index: 571175 (C4 (Web Text))

Max Activating Token Index: 873

Text #12

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.6986. Min Act: -0.1700

Data Index: 1536389 (Python Code)

Max Activating Token Index: 357

Text #13

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.6673. Min Act: -0.1700

Data Index: 1518436 (Python Code)

Max Activating Token Index: 343

Text #14

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.6360. Min Act: -0.1700

Data Index: 801568 (C4 (Web Text))

Max Activating Token Index: 457

Text #15

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.6360. Min Act: -0.1700

Data Index: 199286 (C4 (Web Text))

Max Activating Token Index: 885

Text #16

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.6673. Min Act: -0.1700

Data Index: 1012580 (C4 (Web Text))

Max Activating Token Index: 446

Text #17

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.6673. Min Act: -0.1700

Data Index: 1641236 (Python Code)

Max Activating Token Index: 390

Text #18

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.6360. Min Act: -0.1700

Data Index: 1457729 (Python Code)

Max Activating Token Index: 348

Text #19

Max Range: 3.8550. Min Range: -3.8550

Max Act: 3.6673. Min Act: -0.1700

Data Index: 1537605 (Python Code)

Max Activating Token Index: 939