Model: GELU Model: 2 Layers, 2048 Neurons per Layer

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

Neuron 1180 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-2l')

Text #0

Max Range: 4.3464. Min Range: -4.3464

Max Act: 4.3464. Min Act: -0.1700

Data Index: 428651 (C4 (Web Text))

Max Activating Token Index: 54

Text #1

Max Range: 4.3464. Min Range: -4.3464

Max Act: 4.2292. Min Act: -0.1700

Data Index: 1393529 (Python Code)

Max Activating Token Index: 857

Text #2

Max Range: 4.3464. Min Range: -4.3464

Max Act: 4.2058. Min Act: -0.1700

Data Index: 1423659 (Python Code)

Max Activating Token Index: 804

Text #3

Max Range: 4.3464. Min Range: -4.3464

Max Act: 4.0964. Min Act: -0.1700

Data Index: 740997 (C4 (Web Text))

Max Activating Token Index: 1000

Text #4

Max Range: 4.3464. Min Range: -4.3464

Max Act: 4.0964. Min Act: -0.1700

Data Index: 1551215 (Python Code)

Max Activating Token Index: 1005

Text #5

Max Range: 4.3464. Min Range: -4.3464

Max Act: 4.0651. Min Act: -0.1700

Data Index: 1131822 (C4 (Web Text))

Max Activating Token Index: 937

Text #6

Max Range: 4.3464. Min Range: -4.3464

Max Act: 4.0260. Min Act: -0.1700

Data Index: 1366959 (Python Code)

Max Activating Token Index: 827

Text #7

Max Range: 4.3464. Min Range: -4.3464

Max Act: 3.9948. Min Act: -0.1700

Data Index: 1144529 (C4 (Web Text))

Max Activating Token Index: 543

Text #8

Max Range: 4.3464. Min Range: -4.3464

Max Act: 3.9791. Min Act: -0.1700

Data Index: 1362699 (Python Code)

Max Activating Token Index: 152

Text #9

Max Range: 4.3464. Min Range: -4.3464

Max Act: 4.0182. Min Act: -0.1700

Data Index: 1525821 (Python Code)

Max Activating Token Index: 864

Text #10

Max Range: 4.3464. Min Range: -4.3464

Max Act: 3.9400. Min Act: -0.1700

Data Index: 1180049 (C4 (Web Text))

Max Activating Token Index: 721

Text #11

Max Range: 4.3464. Min Range: -4.3464

Max Act: 3.9088. Min Act: -0.1700

Data Index: 1500662 (Python Code)

Max Activating Token Index: 621

Text #12

Max Range: 4.3464. Min Range: -4.3464

Max Act: 3.9009. Min Act: -0.1699

Data Index: 1541151 (Python Code)

Max Activating Token Index: 506

Text #13

Max Range: 4.3464. Min Range: -4.3464

Max Act: 3.8384. Min Act: -0.1700

Data Index: 486146 (C4 (Web Text))

Max Activating Token Index: 1003

Text #14

Max Range: 4.3464. Min Range: -4.3464

Max Act: 3.8619. Min Act: -0.1700

Data Index: 1526329 (Python Code)

Max Activating Token Index: 59

Text #15

Max Range: 4.3464. Min Range: -4.3464

Max Act: 3.8619. Min Act: -0.1700

Data Index: 66748 (C4 (Web Text))

Max Activating Token Index: 917

Text #16

Max Range: 4.3464. Min Range: -4.3464

Max Act: 3.8619. Min Act: -0.1700

Data Index: 1461698 (Python Code)

Max Activating Token Index: 278

Text #17

Max Range: 4.3464. Min Range: -4.3464

Max Act: 3.8697. Min Act: -0.1700

Data Index: 1262214 (C4 (Web Text))

Max Activating Token Index: 874

Text #18

Max Range: 4.3464. Min Range: -4.3464

Max Act: 3.8540. Min Act: -0.1700

Data Index: 1515856 (Python Code)

Max Activating Token Index: 215

Text #19

Max Range: 4.3464. Min Range: -4.3464

Max Act: 3.8462. Min Act: -0.1700

Data Index: 1072333 (C4 (Web Text))

Max Activating Token Index: 166