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

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

Neuron 850 in Layer 0

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.1012. Min Range: -3.1012

Max Act: 3.1012. Min Act: -0.1700

Data Index: 484659 (C4 (Web Text))

Max Activating Token Index: 698

Text #1

Max Range: 3.1012. Min Range: -3.1012

Max Act: 2.9115. Min Act: -0.1700

Data Index: 1017210 (C4 (Web Text))

Max Activating Token Index: 2

Text #2

Max Range: 3.1012. Min Range: -3.1012

Max Act: 2.8639. Min Act: -0.1700

Data Index: 1523648 (Python Code)

Max Activating Token Index: 5

Text #3

Max Range: 3.1012. Min Range: -3.1012

Max Act: 2.8480. Min Act: -0.1700

Data Index: 1156697 (C4 (Web Text))

Max Activating Token Index: 354

Text #4

Max Range: 3.1012. Min Range: -3.1012

Max Act: 2.8480. Min Act: -0.1700

Data Index: 451590 (C4 (Web Text))

Max Activating Token Index: 236

Text #5

Max Range: 3.1012. Min Range: -3.1012

Max Act: 2.8162. Min Act: -0.1700

Data Index: 199586 (C4 (Web Text))

Max Activating Token Index: 7

Text #6

Max Range: 3.1012. Min Range: -3.1012

Max Act: 2.8162. Min Act: -0.1700

Data Index: 1321300 (C4 (Web Text))

Max Activating Token Index: 327

Text #7

Max Range: 3.1012. Min Range: -3.1012

Max Act: 2.8003. Min Act: -0.1700

Data Index: 1103138 (C4 (Web Text))

Max Activating Token Index: 263

Text #8

Max Range: 3.1012. Min Range: -3.1012

Max Act: 2.7843. Min Act: -0.1700

Data Index: 1129846 (C4 (Web Text))

Max Activating Token Index: 957

Text #9

Max Range: 3.1012. Min Range: -3.1012

Max Act: 2.7843. Min Act: -0.1700

Data Index: 288962 (C4 (Web Text))

Max Activating Token Index: 74

Text #10

Max Range: 3.1012. Min Range: -3.1012

Max Act: 2.7843. Min Act: -0.1700

Data Index: 374710 (C4 (Web Text))

Max Activating Token Index: 639

Text #11

Max Range: 3.1012. Min Range: -3.1012

Max Act: 2.7843. Min Act: -0.1700

Data Index: 953846 (C4 (Web Text))

Max Activating Token Index: 993

Text #12

Max Range: 3.1012. Min Range: -3.1012

Max Act: 2.7684. Min Act: -0.1700

Data Index: 1428586 (Python Code)

Max Activating Token Index: 77

Text #13

Max Range: 3.1012. Min Range: -3.1012

Max Act: 2.7684. Min Act: -0.1700

Data Index: 406855 (C4 (Web Text))

Max Activating Token Index: 164

Text #14

Max Range: 3.1012. Min Range: -3.1012

Max Act: 2.7684. Min Act: -0.1700

Data Index: 1311876 (C4 (Web Text))

Max Activating Token Index: 127

Text #15

Max Range: 3.1012. Min Range: -3.1012

Max Act: 2.7684. Min Act: -0.1700

Data Index: 1293504 (C4 (Web Text))

Max Activating Token Index: 674

Text #16

Max Range: 3.1012. Min Range: -3.1012

Max Act: 2.7843. Min Act: -0.1700

Data Index: 692167 (C4 (Web Text))

Max Activating Token Index: 455

Text #17

Max Range: 3.1012. Min Range: -3.1012

Max Act: 2.7684. Min Act: -0.1700

Data Index: 1045089 (C4 (Web Text))

Max Activating Token Index: 730

Text #18

Max Range: 3.1012. Min Range: -3.1012

Max Act: 2.7843. Min Act: -0.1700

Data Index: 176206 (C4 (Web Text))

Max Activating Token Index: 996

Text #19

Max Range: 3.1012. Min Range: -3.1012

Max Act: 2.7365. Min Act: -0.1700

Data Index: 272731 (C4 (Web Text))

Max Activating Token Index: 893