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

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

Neuron 1045 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: 3.6033. Min Range: -3.6033

Max Act: 3.6033. Min Act: -0.1700

Data Index: 525498 (C4 (Web Text))

Max Activating Token Index: 764

Text #1

Max Range: 3.6033. Min Range: -3.6033

Max Act: 3.4624. Min Act: -0.1700

Data Index: 870713 (C4 (Web Text))

Max Activating Token Index: 96

Text #2

Max Range: 3.6033. Min Range: -3.6033

Max Act: 3.3996. Min Act: -0.1700

Data Index: 1600572 (Python Code)

Max Activating Token Index: 151

Text #3

Max Range: 3.6033. Min Range: -3.6033

Max Act: 3.4153. Min Act: -0.1700

Data Index: 1525472 (Python Code)

Max Activating Token Index: 103

Text #4

Max Range: 3.6033. Min Range: -3.6033

Max Act: 3.3683. Min Act: -0.1700

Data Index: 1453268 (Python Code)

Max Activating Token Index: 490

Text #5

Max Range: 3.6033. Min Range: -3.6033

Max Act: 3.3683. Min Act: -0.1699

Data Index: 206744 (C4 (Web Text))

Max Activating Token Index: 147

Text #6

Max Range: 3.6033. Min Range: -3.6033

Max Act: 3.3996. Min Act: -0.1700

Data Index: 356964 (C4 (Web Text))

Max Activating Token Index: 485

Text #7

Max Range: 3.6033. Min Range: -3.6033

Max Act: 3.3369. Min Act: -0.1700

Data Index: 1532186 (Python Code)

Max Activating Token Index: 146

Text #8

Max Range: 3.6033. Min Range: -3.6033

Max Act: 3.3369. Min Act: -0.1700

Data Index: 484619 (C4 (Web Text))

Max Activating Token Index: 910

Text #9

Max Range: 3.6033. Min Range: -3.6033

Max Act: 3.3055. Min Act: -0.1700

Data Index: 411585 (C4 (Web Text))

Max Activating Token Index: 692

Text #10

Max Range: 3.6033. Min Range: -3.6033

Max Act: 3.3212. Min Act: -0.1700

Data Index: 954457 (C4 (Web Text))

Max Activating Token Index: 565

Text #11

Max Range: 3.6033. Min Range: -3.6033

Max Act: 3.3212. Min Act: -0.1700

Data Index: 290132 (C4 (Web Text))

Max Activating Token Index: 823

Text #12

Max Range: 3.6033. Min Range: -3.6033

Max Act: 3.3055. Min Act: -0.1700

Data Index: 1516714 (Python Code)

Max Activating Token Index: 218

Text #13

Max Range: 3.6033. Min Range: -3.6033

Max Act: 3.2741. Min Act: -0.1700

Data Index: 1385878 (Python Code)

Max Activating Token Index: 455

Text #14

Max Range: 3.6033. Min Range: -3.6033

Max Act: 3.3369. Min Act: -0.1700

Data Index: 66997 (C4 (Web Text))

Max Activating Token Index: 427

Text #15

Max Range: 3.6033. Min Range: -3.6033

Max Act: 3.2898. Min Act: -0.1700

Data Index: 274547 (C4 (Web Text))

Max Activating Token Index: 462

Text #16

Max Range: 3.6033. Min Range: -3.6033

Max Act: 3.3212. Min Act: -0.1700

Data Index: 1476341 (Python Code)

Max Activating Token Index: 168

Text #17

Max Range: 3.6033. Min Range: -3.6033

Max Act: 3.2898. Min Act: -0.1700

Data Index: 1509322 (Python Code)

Max Activating Token Index: 143

Text #18

Max Range: 3.6033. Min Range: -3.6033

Max Act: 3.2898. Min Act: -0.1700

Data Index: 1618649 (Python Code)

Max Activating Token Index: 368

Text #19

Max Range: 3.6033. Min Range: -3.6033

Max Act: 3.2898. Min Act: -0.1700

Data Index: 1566838 (Python Code)

Max Activating Token Index: 281