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

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

Neuron 2004 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.4800. Min Act: -0.1700

Data Index: 1277061 (C4 (Web Text))

Max Activating Token Index: 1012

Text #1

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.4175. Min Act: -0.1700

Data Index: 1111398 (C4 (Web Text))

Max Activating Token Index: 258

Text #2

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.3550. Min Act: -0.1700

Data Index: 680252 (C4 (Web Text))

Max Activating Token Index: 203

Text #3

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.3238. Min Act: -0.1700

Data Index: 1299224 (C4 (Web Text))

Max Activating Token Index: 1018

Text #4

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.3238. Min Act: -0.1700

Data Index: 881454 (C4 (Web Text))

Max Activating Token Index: 979

Text #5

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.2925. Min Act: -0.1700

Data Index: 423639 (C4 (Web Text))

Max Activating Token Index: 105

Text #6

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.3238. Min Act: -0.1700

Data Index: 475910 (C4 (Web Text))

Max Activating Token Index: 156

Text #7

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.2613. Min Act: -0.1700

Data Index: 1120699 (C4 (Web Text))

Max Activating Token Index: 964

Text #8

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.2300. Min Act: -0.1700

Data Index: 867802 (C4 (Web Text))

Max Activating Token Index: 973

Text #9

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.1988. Min Act: -0.1700

Data Index: 132469 (C4 (Web Text))

Max Activating Token Index: 948

Text #10

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.2300. Min Act: -0.1700

Data Index: 907940 (C4 (Web Text))

Max Activating Token Index: 126

Text #11

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.1988. Min Act: -0.1700

Data Index: 153064 (C4 (Web Text))

Max Activating Token Index: 838

Text #12

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.1050. Min Act: -0.1700

Data Index: 1246604 (C4 (Web Text))

Max Activating Token Index: 232

Text #13

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.1988. Min Act: -0.1700

Data Index: 305885 (C4 (Web Text))

Max Activating Token Index: 982

Text #14

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.1363. Min Act: -0.1700

Data Index: 35890 (C4 (Web Text))

Max Activating Token Index: 130

Text #15

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.1675. Min Act: -0.1700

Data Index: 355953 (C4 (Web Text))

Max Activating Token Index: 336

Text #16

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.1363. Min Act: -0.1700

Data Index: 414572 (C4 (Web Text))

Max Activating Token Index: 183

Text #17

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.1675. Min Act: -0.1700

Data Index: 781676 (C4 (Web Text))

Max Activating Token Index: 183

Text #18

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.0738. Min Act: -0.1700

Data Index: 927172 (C4 (Web Text))

Max Activating Token Index: 292

Text #19

Max Range: 5.4800. Min Range: -5.4800

Max Act: 5.0738. Min Act: -0.1700

Data Index: 1229362 (C4 (Web Text))

Max Activating Token Index: 602