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

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

Neuron 1974 in Layer 0

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.4764. Min Act: -0.1700

Data Index: 1107485 (C4 (Web Text))

Max Activating Token Index: 533

Text #1

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.4608. Min Act: -0.1700

Data Index: 673913 (C4 (Web Text))

Max Activating Token Index: 970

Text #2

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.4608. Min Act: -0.1700

Data Index: 662856 (C4 (Web Text))

Max Activating Token Index: 640

Text #3

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.3823. Min Act: -0.1700

Data Index: 1351574 (C4 (Web Text))

Max Activating Token Index: 421

Text #4

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.3980. Min Act: -0.1700

Data Index: 162295 (C4 (Web Text))

Max Activating Token Index: 780

Text #5

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.3823. Min Act: -0.1700

Data Index: 563038 (C4 (Web Text))

Max Activating Token Index: 189

Text #6

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.3823. Min Act: -0.1700

Data Index: 862829 (C4 (Web Text))

Max Activating Token Index: 579

Text #7

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.3667. Min Act: -0.1700

Data Index: 604876 (C4 (Web Text))

Max Activating Token Index: 903

Text #8

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.3510. Min Act: -0.1700

Data Index: 1258890 (C4 (Web Text))

Max Activating Token Index: 166

Text #9

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.3510. Min Act: -0.1700

Data Index: 1122878 (C4 (Web Text))

Max Activating Token Index: 576

Text #10

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.3510. Min Act: -0.1700

Data Index: 841052 (C4 (Web Text))

Max Activating Token Index: 891

Text #11

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.3353. Min Act: -0.1700

Data Index: 373146 (C4 (Web Text))

Max Activating Token Index: 891

Text #12

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.3353. Min Act: -0.1700

Data Index: 1237409 (C4 (Web Text))

Max Activating Token Index: 356

Text #13

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.3353. Min Act: -0.1700

Data Index: 825048 (C4 (Web Text))

Max Activating Token Index: 461

Text #14

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.3196. Min Act: -0.1700

Data Index: 933427 (C4 (Web Text))

Max Activating Token Index: 911

Text #15

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.3039. Min Act: -0.1700

Data Index: 898519 (C4 (Web Text))

Max Activating Token Index: 318

Text #16

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.3196. Min Act: -0.1700

Data Index: 694694 (C4 (Web Text))

Max Activating Token Index: 260

Text #17

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.3196. Min Act: -0.1700

Data Index: 866412 (C4 (Web Text))

Max Activating Token Index: 810

Text #18

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.3039. Min Act: -0.1700

Data Index: 1506010 (Python Code)

Max Activating Token Index: 316

Text #19

Max Range: 3.4764. Min Range: -3.4764

Max Act: 3.2882. Min Act: -0.1700

Data Index: 134537 (C4 (Web Text))

Max Activating Token Index: 756