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

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

Neuron 960 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-4l')

Text #0

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.4105. Min Act: -0.1700

Data Index: 296039 (C4 (Web Text))

Max Activating Token Index: 640

Text #1

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.3791. Min Act: -0.1700

Data Index: 401218 (C4 (Web Text))

Max Activating Token Index: 659

Text #2

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.3320. Min Act: -0.1700

Data Index: 1571895 (Python Code)

Max Activating Token Index: 438

Text #3

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.3163. Min Act: -0.1700

Data Index: 918564 (C4 (Web Text))

Max Activating Token Index: 513

Text #4

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.3320. Min Act: -0.1700

Data Index: 1153171 (C4 (Web Text))

Max Activating Token Index: 1023

Text #5

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.2692. Min Act: -0.1700

Data Index: 1416602 (Python Code)

Max Activating Token Index: 113

Text #6

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.2535. Min Act: -0.1700

Data Index: 425394 (C4 (Web Text))

Max Activating Token Index: 728

Text #7

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.2535. Min Act: -0.1700

Data Index: 953583 (C4 (Web Text))

Max Activating Token Index: 753

Text #8

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.2377. Min Act: -0.1700

Data Index: 1527801 (Python Code)

Max Activating Token Index: 796

Text #9

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.1906. Min Act: -0.1700

Data Index: 1165853 (C4 (Web Text))

Max Activating Token Index: 42

Text #10

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.2063. Min Act: -0.1700

Data Index: 272828 (C4 (Web Text))

Max Activating Token Index: 859

Text #11

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.2220. Min Act: -0.1700

Data Index: 122152 (C4 (Web Text))

Max Activating Token Index: 990

Text #12

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.1906. Min Act: -0.1700

Data Index: 628848 (C4 (Web Text))

Max Activating Token Index: 642

Text #13

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.1906. Min Act: -0.1700

Data Index: 675132 (C4 (Web Text))

Max Activating Token Index: 575

Text #14

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.1906. Min Act: -0.1700

Data Index: 1509860 (Python Code)

Max Activating Token Index: 356

Text #15

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.1906. Min Act: -0.1700

Data Index: 170159 (C4 (Web Text))

Max Activating Token Index: 971

Text #16

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.1748. Min Act: -0.1700

Data Index: 788772 (C4 (Web Text))

Max Activating Token Index: 364

Text #17

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.1591. Min Act: -0.1700

Data Index: 1198326 (C4 (Web Text))

Max Activating Token Index: 759

Text #18

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.1906. Min Act: -0.1700

Data Index: 526904 (C4 (Web Text))

Max Activating Token Index: 1023

Text #19

Max Range: 3.4105. Min Range: -3.4105

Max Act: 3.1433. Min Act: -0.1700

Data Index: 1478363 (Python Code)

Max Activating Token Index: 693