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

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

Neuron 1256 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-3l')

Text #0

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.3865. Min Act: -0.1700

Data Index: 1612868 (Python Code)

Max Activating Token Index: 898

Text #1

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.3080. Min Act: -0.1700

Data Index: 1535669 (Python Code)

Max Activating Token Index: 893

Text #2

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.3080. Min Act: -0.1700

Data Index: 1052321 (C4 (Web Text))

Max Activating Token Index: 971

Text #3

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.2452. Min Act: -0.1700

Data Index: 1647367 (Python Code)

Max Activating Token Index: 812

Text #4

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.2452. Min Act: -0.1700

Data Index: 1652684 (Python Code)

Max Activating Token Index: 808

Text #5

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.2137. Min Act: -0.1700

Data Index: 890007 (C4 (Web Text))

Max Activating Token Index: 1020

Text #6

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.1665. Min Act: -0.1700

Data Index: 1531875 (Python Code)

Max Activating Token Index: 1008

Text #7

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.2295. Min Act: -0.1700

Data Index: 1575383 (Python Code)

Max Activating Token Index: 869

Text #8

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.1665. Min Act: -0.1700

Data Index: 1621915 (Python Code)

Max Activating Token Index: 1010

Text #9

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.2137. Min Act: -0.1700

Data Index: 1464284 (Python Code)

Max Activating Token Index: 893

Text #10

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.1665. Min Act: -0.1700

Data Index: 1483400 (Python Code)

Max Activating Token Index: 1022

Text #11

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.0877. Min Act: -0.1700

Data Index: 1406707 (Python Code)

Max Activating Token Index: 902

Text #12

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.0877. Min Act: -0.1700

Data Index: 1368412 (Python Code)

Max Activating Token Index: 968

Text #13

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.1035. Min Act: -0.1700

Data Index: 1381728 (Python Code)

Max Activating Token Index: 840

Text #14

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.1193. Min Act: -0.1700

Data Index: 186205 (C4 (Web Text))

Max Activating Token Index: 958

Text #15

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.0562. Min Act: -0.1700

Data Index: 1452421 (Python Code)

Max Activating Token Index: 891

Text #16

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.0562. Min Act: -0.1700

Data Index: 1468648 (Python Code)

Max Activating Token Index: 936

Text #17

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.0720. Min Act: -0.1700

Data Index: 1430036 (Python Code)

Max Activating Token Index: 864

Text #18

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.0404. Min Act: -0.1700

Data Index: 1378397 (Python Code)

Max Activating Token Index: 898

Text #19

Max Range: 3.3865. Min Range: -3.3865

Max Act: 3.0562. Min Act: -0.1700

Data Index: 1400933 (Python Code)

Max Activating Token Index: 977