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

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

Neuron 2036 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 4.1748. Min Range: -4.1748

Max Act: 4.1748. Min Act: -0.1700

Data Index: 1206984 (C4 (Web Text))

Max Activating Token Index: 196

Text #1

Max Range: 4.1748. Min Range: -4.1748

Max Act: 3.9872. Min Act: -0.1700

Data Index: 107567 (C4 (Web Text))

Max Activating Token Index: 810

Text #2

Max Range: 4.1748. Min Range: -4.1748

Max Act: 3.9560. Min Act: -0.1700

Data Index: 1175699 (C4 (Web Text))

Max Activating Token Index: 494

Text #3

Max Range: 4.1748. Min Range: -4.1748

Max Act: 3.9872. Min Act: -0.1700

Data Index: 1111868 (C4 (Web Text))

Max Activating Token Index: 230

Text #4

Max Range: 4.1748. Min Range: -4.1748

Max Act: 3.9872. Min Act: -0.1700

Data Index: 1067171 (C4 (Web Text))

Max Activating Token Index: 202

Text #5

Max Range: 4.1748. Min Range: -4.1748

Max Act: 3.9560. Min Act: -0.1700

Data Index: 1627370 (Python Code)

Max Activating Token Index: 13

Text #6

Max Range: 4.1748. Min Range: -4.1748

Max Act: 3.9247. Min Act: -0.1700

Data Index: 503148 (C4 (Web Text))

Max Activating Token Index: 196

Text #7

Max Range: 4.1748. Min Range: -4.1748

Max Act: 3.9560. Min Act: -0.1700

Data Index: 732414 (C4 (Web Text))

Max Activating Token Index: 183

Text #8

Max Range: 4.1748. Min Range: -4.1748

Max Act: 3.9560. Min Act: -0.1700

Data Index: 35461 (C4 (Web Text))

Max Activating Token Index: 508

Text #9

Max Range: 4.1748. Min Range: -4.1748

Max Act: 3.8934. Min Act: -0.1700

Data Index: 598050 (C4 (Web Text))

Max Activating Token Index: 152

Text #10

Max Range: 4.1748. Min Range: -4.1748

Max Act: 3.9247. Min Act: -0.1700

Data Index: 1386040 (Python Code)

Max Activating Token Index: 13

Text #11

Max Range: 4.1748. Min Range: -4.1748

Max Act: 3.9560. Min Act: -0.1700

Data Index: 1334899 (C4 (Web Text))

Max Activating Token Index: 823

Text #12

Max Range: 4.1748. Min Range: -4.1748

Max Act: 3.9247. Min Act: -0.1700

Data Index: 307519 (C4 (Web Text))

Max Activating Token Index: 394

Text #13

Max Range: 4.1748. Min Range: -4.1748

Max Act: 3.8934. Min Act: -0.1700

Data Index: 769366 (C4 (Web Text))

Max Activating Token Index: 270

Text #14

Max Range: 4.1748. Min Range: -4.1748

Max Act: 3.9560. Min Act: -0.1700

Data Index: 205206 (C4 (Web Text))

Max Activating Token Index: 609

Text #15

Max Range: 4.1748. Min Range: -4.1748

Max Act: 3.9560. Min Act: -0.1700

Data Index: 282227 (C4 (Web Text))

Max Activating Token Index: 121

Text #16

Max Range: 4.1748. Min Range: -4.1748

Max Act: 3.9247. Min Act: -0.1700

Data Index: 210152 (C4 (Web Text))

Max Activating Token Index: 799

Text #17

Max Range: 4.1748. Min Range: -4.1748

Max Act: 3.9247. Min Act: -0.1700

Data Index: 401463 (C4 (Web Text))

Max Activating Token Index: 371

Text #18

Max Range: 4.1748. Min Range: -4.1748

Max Act: 3.8934. Min Act: -0.1700

Data Index: 981154 (C4 (Web Text))

Max Activating Token Index: 929

Text #19

Max Range: 4.1748. Min Range: -4.1748

Max Act: 3.9247. Min Act: -0.1700

Data Index: 717003 (C4 (Web Text))

Max Activating Token Index: 357