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

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

Neuron 1287 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: 7.1775. Min Range: -7.1775

Max Act: 7.1463. Min Act: -0.1700

Data Index: 906995 (C4 (Web Text))

Max Activating Token Index: 896

Text #1

Max Range: 7.1775. Min Range: -7.1775

Max Act: 7.1775. Min Act: -0.1700

Data Index: 1359855 (Python Code)

Max Activating Token Index: 976

Text #2

Max Range: 7.1775. Min Range: -7.1775

Max Act: 7.1150. Min Act: -0.1700

Data Index: 595814 (C4 (Web Text))

Max Activating Token Index: 532

Text #3

Max Range: 7.1775. Min Range: -7.1775

Max Act: 7.0525. Min Act: -0.1700

Data Index: 395172 (C4 (Web Text))

Max Activating Token Index: 951

Text #4

Max Range: 7.1775. Min Range: -7.1775

Max Act: 7.0213. Min Act: -0.1700

Data Index: 98732 (C4 (Web Text))

Max Activating Token Index: 772

Text #5

Max Range: 7.1775. Min Range: -7.1775

Max Act: 6.9588. Min Act: -0.1700

Data Index: 1411466 (Python Code)

Max Activating Token Index: 79

Text #6

Max Range: 7.1775. Min Range: -7.1775

Max Act: 6.8963. Min Act: -0.1700

Data Index: 1305879 (C4 (Web Text))

Max Activating Token Index: 348

Text #7

Max Range: 7.1775. Min Range: -7.1775

Max Act: 6.7713. Min Act: -0.1700

Data Index: 1420487 (Python Code)

Max Activating Token Index: 871

Text #8

Max Range: 7.1775. Min Range: -7.1775

Max Act: 6.8025. Min Act: -0.1700

Data Index: 690207 (C4 (Web Text))

Max Activating Token Index: 559

Text #9

Max Range: 7.1775. Min Range: -7.1775

Max Act: 6.8025. Min Act: -0.1700

Data Index: 517274 (C4 (Web Text))

Max Activating Token Index: 341

Text #10

Max Range: 7.1775. Min Range: -7.1775

Max Act: 6.8025. Min Act: -0.1700

Data Index: 422351 (C4 (Web Text))

Max Activating Token Index: 356

Text #11

Max Range: 7.1775. Min Range: -7.1775

Max Act: 6.7088. Min Act: -0.1700

Data Index: 1580983 (Python Code)

Max Activating Token Index: 764

Text #12

Max Range: 7.1775. Min Range: -7.1775

Max Act: 6.7713. Min Act: -0.1700

Data Index: 10551 (C4 (Web Text))

Max Activating Token Index: 492

Text #13

Max Range: 7.1775. Min Range: -7.1775

Max Act: 6.7400. Min Act: -0.1700

Data Index: 1186095 (C4 (Web Text))

Max Activating Token Index: 643

Text #14

Max Range: 7.1775. Min Range: -7.1775

Max Act: 6.6775. Min Act: -0.1700

Data Index: 1543349 (Python Code)

Max Activating Token Index: 723

Text #15

Max Range: 7.1775. Min Range: -7.1775

Max Act: 6.6150. Min Act: -0.1700

Data Index: 302023 (C4 (Web Text))

Max Activating Token Index: 876

Text #16

Max Range: 7.1775. Min Range: -7.1775

Max Act: 6.6775. Min Act: -0.1700

Data Index: 1174826 (C4 (Web Text))

Max Activating Token Index: 972

Text #17

Max Range: 7.1775. Min Range: -7.1775

Max Act: 6.6463. Min Act: -0.1700

Data Index: 314571 (C4 (Web Text))

Max Activating Token Index: 920

Text #18

Max Range: 7.1775. Min Range: -7.1775

Max Act: 6.6463. Min Act: -0.1700

Data Index: 1564580 (Python Code)

Max Activating Token Index: 876

Text #19

Max Range: 7.1775. Min Range: -7.1775

Max Act: 6.6775. Min Act: -0.1700

Data Index: 904075 (C4 (Web Text))

Max Activating Token Index: 887