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

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

Neuron 1678 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: 4.6955. Min Range: -4.6955

Max Act: 4.6955. Min Act: -0.1700

Data Index: 250762 (C4 (Web Text))

Max Activating Token Index: 372

Text #1

Max Range: 4.6955. Min Range: -4.6955

Max Act: 3.8516. Min Act: -0.1700

Data Index: 1000782 (C4 (Web Text))

Max Activating Token Index: 925

Text #2

Max Range: 4.6955. Min Range: -4.6955

Max Act: 3.8203. Min Act: -0.1700

Data Index: 498683 (C4 (Web Text))

Max Activating Token Index: 978

Text #3

Max Range: 4.6955. Min Range: -4.6955

Max Act: 3.8203. Min Act: -0.1700

Data Index: 329743 (C4 (Web Text))

Max Activating Token Index: 916

Text #4

Max Range: 4.6955. Min Range: -4.6955

Max Act: 3.7890. Min Act: -0.1700

Data Index: 342940 (C4 (Web Text))

Max Activating Token Index: 842

Text #5

Max Range: 4.6955. Min Range: -4.6955

Max Act: 3.7890. Min Act: -0.1700

Data Index: 434829 (C4 (Web Text))

Max Activating Token Index: 985

Text #6

Max Range: 4.6955. Min Range: -4.6955

Max Act: 3.7890. Min Act: -0.1700

Data Index: 1150339 (C4 (Web Text))

Max Activating Token Index: 739

Text #7

Max Range: 4.6955. Min Range: -4.6955

Max Act: 3.7890. Min Act: -0.1700

Data Index: 886009 (C4 (Web Text))

Max Activating Token Index: 925

Text #8

Max Range: 4.6955. Min Range: -4.6955

Max Act: 3.7577. Min Act: -0.1700

Data Index: 886636 (C4 (Web Text))

Max Activating Token Index: 718

Text #9

Max Range: 4.6955. Min Range: -4.6955

Max Act: 3.7890. Min Act: -0.1700

Data Index: 1267923 (C4 (Web Text))

Max Activating Token Index: 773

Text #10

Max Range: 4.6955. Min Range: -4.6955

Max Act: 3.7890. Min Act: -0.1700

Data Index: 976902 (C4 (Web Text))

Max Activating Token Index: 995

Text #11

Max Range: 4.6955. Min Range: -4.6955

Max Act: 3.7890. Min Act: -0.1700

Data Index: 311223 (C4 (Web Text))

Max Activating Token Index: 917

Text #12

Max Range: 4.6955. Min Range: -4.6955

Max Act: 3.7890. Min Act: -0.1700

Data Index: 26659 (C4 (Web Text))

Max Activating Token Index: 915

Text #13

Max Range: 4.6955. Min Range: -4.6955

Max Act: 3.7577. Min Act: -0.1700

Data Index: 708385 (C4 (Web Text))

Max Activating Token Index: 780

Text #14

Max Range: 4.6955. Min Range: -4.6955

Max Act: 3.7890. Min Act: -0.1699

Data Index: 753235 (C4 (Web Text))

Max Activating Token Index: 875

Text #15

Max Range: 4.6955. Min Range: -4.6955

Max Act: 3.7890. Min Act: -0.1700

Data Index: 115889 (C4 (Web Text))

Max Activating Token Index: 766

Text #16

Max Range: 4.6955. Min Range: -4.6955

Max Act: 3.7577. Min Act: -0.1700

Data Index: 235579 (C4 (Web Text))

Max Activating Token Index: 739

Text #17

Max Range: 4.6955. Min Range: -4.6955

Max Act: 3.7264. Min Act: -0.1700

Data Index: 16797 (C4 (Web Text))

Max Activating Token Index: 1022

Text #18

Max Range: 4.6955. Min Range: -4.6955

Max Act: 3.7577. Min Act: -0.1700

Data Index: 946994 (C4 (Web Text))

Max Activating Token Index: 581

Text #19

Max Range: 4.6955. Min Range: -4.6955

Max Act: 3.7264. Min Act: -0.1700

Data Index: 324685 (C4 (Web Text))

Max Activating Token Index: 746