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

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

Neuron 695 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: 7.0550. Min Range: -7.0550

Max Act: 7.0550. Min Act: -0.1700

Data Index: 186262 (C4 (Web Text))

Max Activating Token Index: 936

Text #1

Max Range: 7.0550. Min Range: -7.0550

Max Act: 6.6800. Min Act: -0.1700

Data Index: 399709 (C4 (Web Text))

Max Activating Token Index: 309

Text #2

Max Range: 7.0550. Min Range: -7.0550

Max Act: 6.6487. Min Act: -0.1700

Data Index: 48477 (C4 (Web Text))

Max Activating Token Index: 214

Text #3

Max Range: 7.0550. Min Range: -7.0550

Max Act: 6.5237. Min Act: -0.1700

Data Index: 736427 (C4 (Web Text))

Max Activating Token Index: 275

Text #4

Max Range: 7.0550. Min Range: -7.0550

Max Act: 6.3987. Min Act: -0.1700

Data Index: 730850 (C4 (Web Text))

Max Activating Token Index: 492

Text #5

Max Range: 7.0550. Min Range: -7.0550

Max Act: 6.3987. Min Act: -0.1700

Data Index: 481718 (C4 (Web Text))

Max Activating Token Index: 832

Text #6

Max Range: 7.0550. Min Range: -7.0550

Max Act: 6.4300. Min Act: -0.1700

Data Index: 1256331 (C4 (Web Text))

Max Activating Token Index: 459

Text #7

Max Range: 7.0550. Min Range: -7.0550

Max Act: 6.2737. Min Act: -0.1700

Data Index: 945551 (C4 (Web Text))

Max Activating Token Index: 104

Text #8

Max Range: 7.0550. Min Range: -7.0550

Max Act: 6.3362. Min Act: -0.1700

Data Index: 1265883 (C4 (Web Text))

Max Activating Token Index: 785

Text #9

Max Range: 7.0550. Min Range: -7.0550

Max Act: 6.3362. Min Act: -0.1700

Data Index: 1116803 (C4 (Web Text))

Max Activating Token Index: 247

Text #10

Max Range: 7.0550. Min Range: -7.0550

Max Act: 6.2737. Min Act: -0.1700

Data Index: 598988 (C4 (Web Text))

Max Activating Token Index: 262

Text #11

Max Range: 7.0550. Min Range: -7.0550

Max Act: 6.1800. Min Act: -0.1700

Data Index: 1272928 (C4 (Web Text))

Max Activating Token Index: 388

Text #12

Max Range: 7.0550. Min Range: -7.0550

Max Act: 6.2112. Min Act: -0.1699

Data Index: 1423728 (Python Code)

Max Activating Token Index: 507

Text #13

Max Range: 7.0550. Min Range: -7.0550

Max Act: 6.2112. Min Act: -0.1700

Data Index: 767234 (C4 (Web Text))

Max Activating Token Index: 304

Text #14

Max Range: 7.0550. Min Range: -7.0550

Max Act: 6.2112. Min Act: -0.1700

Data Index: 1204535 (C4 (Web Text))

Max Activating Token Index: 139

Text #15

Max Range: 7.0550. Min Range: -7.0550

Max Act: 6.2112. Min Act: -0.1700

Data Index: 279661 (C4 (Web Text))

Max Activating Token Index: 719

Text #16

Max Range: 7.0550. Min Range: -7.0550

Max Act: 6.1800. Min Act: -0.1700

Data Index: 882740 (C4 (Web Text))

Max Activating Token Index: 332

Text #17

Max Range: 7.0550. Min Range: -7.0550

Max Act: 6.1487. Min Act: -0.1700

Data Index: 1121763 (C4 (Web Text))

Max Activating Token Index: 702

Text #18

Max Range: 7.0550. Min Range: -7.0550

Max Act: 6.1800. Min Act: -0.1700

Data Index: 623526 (C4 (Web Text))

Max Activating Token Index: 795

Text #19

Max Range: 7.0550. Min Range: -7.0550

Max Act: 6.1487. Min Act: -0.1700

Data Index: 430131 (C4 (Web Text))

Max Activating Token Index: 490