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

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

Neuron 1873 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: 7.3998. Min Range: -7.3998

Max Act: 7.3998. Min Act: -0.1700

Data Index: 858197 (C4 (Web Text))

Max Activating Token Index: 948

Text #1

Max Range: 7.3998. Min Range: -7.3998

Max Act: 7.1967. Min Act: -0.1700

Data Index: 1421663 (Python Code)

Max Activating Token Index: 610

Text #2

Max Range: 7.3998. Min Range: -7.3998

Max Act: 7.2279. Min Act: -0.1700

Data Index: 1359856 (Python Code)

Max Activating Token Index: 712

Text #3

Max Range: 7.3998. Min Range: -7.3998

Max Act: 7.1654. Min Act: -0.1700

Data Index: 1483950 (Python Code)

Max Activating Token Index: 890

Text #4

Max Range: 7.3998. Min Range: -7.3998

Max Act: 7.0873. Min Act: -0.1700

Data Index: 345042 (C4 (Web Text))

Max Activating Token Index: 296

Text #5

Max Range: 7.3998. Min Range: -7.3998

Max Act: 7.0092. Min Act: -0.1700

Data Index: 1641864 (Python Code)

Max Activating Token Index: 335

Text #6

Max Range: 7.3998. Min Range: -7.3998

Max Act: 7.0404. Min Act: -0.1700

Data Index: 1569315 (Python Code)

Max Activating Token Index: 668

Text #7

Max Range: 7.3998. Min Range: -7.3998

Max Act: 7.1186. Min Act: -0.1700

Data Index: 380782 (C4 (Web Text))

Max Activating Token Index: 941

Text #8

Max Range: 7.3998. Min Range: -7.3998

Max Act: 6.9311. Min Act: -0.1700

Data Index: 412656 (C4 (Web Text))

Max Activating Token Index: 1014

Text #9

Max Range: 7.3998. Min Range: -7.3998

Max Act: 6.7748. Min Act: -0.1700

Data Index: 1574101 (Python Code)

Max Activating Token Index: 496

Text #10

Max Range: 7.3998. Min Range: -7.3998

Max Act: 6.7904. Min Act: -0.1700

Data Index: 897614 (C4 (Web Text))

Max Activating Token Index: 345

Text #11

Max Range: 7.3998. Min Range: -7.3998

Max Act: 6.8373. Min Act: -0.1700

Data Index: 1576246 (Python Code)

Max Activating Token Index: 468

Text #12

Max Range: 7.3998. Min Range: -7.3998

Max Act: 6.8061. Min Act: -0.1700

Data Index: 1546110 (Python Code)

Max Activating Token Index: 232

Text #13

Max Range: 7.3998. Min Range: -7.3998

Max Act: 6.7904. Min Act: -0.1700

Data Index: 834256 (C4 (Web Text))

Max Activating Token Index: 972

Text #14

Max Range: 7.3998. Min Range: -7.3998

Max Act: 6.8686. Min Act: -0.1700

Data Index: 1047987 (C4 (Web Text))

Max Activating Token Index: 1016

Text #15

Max Range: 7.3998. Min Range: -7.3998

Max Act: 6.7748. Min Act: -0.1700

Data Index: 915232 (C4 (Web Text))

Max Activating Token Index: 276

Text #16

Max Range: 7.3998. Min Range: -7.3998

Max Act: 6.7123. Min Act: -0.1700

Data Index: 554427 (C4 (Web Text))

Max Activating Token Index: 434

Text #17

Max Range: 7.3998. Min Range: -7.3998

Max Act: 6.6498. Min Act: -0.1700

Data Index: 1561984 (Python Code)

Max Activating Token Index: 481

Text #18

Max Range: 7.3998. Min Range: -7.3998

Max Act: 6.6811. Min Act: -0.1700

Data Index: 1083321 (C4 (Web Text))

Max Activating Token Index: 279

Text #19

Max Range: 7.3998. Min Range: -7.3998

Max Act: 6.7436. Min Act: -0.1700

Data Index: 1040689 (C4 (Web Text))

Max Activating Token Index: 1009