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

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

Neuron 485 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: 7.8363. Min Range: -7.8363

Max Act: 7.8363. Min Act: -0.1700

Data Index: 295228 (C4 (Web Text))

Max Activating Token Index: 442

Text #1

Max Range: 7.8363. Min Range: -7.8363

Max Act: 7.5863. Min Act: -0.1700

Data Index: 128453 (C4 (Web Text))

Max Activating Token Index: 886

Text #2

Max Range: 7.8363. Min Range: -7.8363

Max Act: 7.6488. Min Act: -0.1700

Data Index: 977983 (C4 (Web Text))

Max Activating Token Index: 645

Text #3

Max Range: 7.8363. Min Range: -7.8363

Max Act: 7.3988. Min Act: -0.1700

Data Index: 1353281 (C4 (Web Text))

Max Activating Token Index: 249

Text #4

Max Range: 7.8363. Min Range: -7.8363

Max Act: 7.5238. Min Act: -0.1690

Data Index: 1352075 (C4 (Web Text))

Max Activating Token Index: 161

Text #5

Max Range: 7.8363. Min Range: -7.8363

Max Act: 7.4613. Min Act: -0.1699

Data Index: 540696 (C4 (Web Text))

Max Activating Token Index: 955

Text #6

Max Range: 7.8363. Min Range: -7.8363

Max Act: 7.4300. Min Act: -0.1700

Data Index: 459245 (C4 (Web Text))

Max Activating Token Index: 373

Text #7

Max Range: 7.8363. Min Range: -7.8363

Max Act: 7.3363. Min Act: -0.1700

Data Index: 144434 (C4 (Web Text))

Max Activating Token Index: 873

Text #8

Max Range: 7.8363. Min Range: -7.8363

Max Act: 7.3363. Min Act: -0.1700

Data Index: 1164094 (C4 (Web Text))

Max Activating Token Index: 597

Text #9

Max Range: 7.8363. Min Range: -7.8363

Max Act: 7.2425. Min Act: -0.1697

Data Index: 47596 (C4 (Web Text))

Max Activating Token Index: 918

Text #10

Max Range: 7.8363. Min Range: -7.8363

Max Act: 7.3675. Min Act: -0.1700

Data Index: 1156787 (C4 (Web Text))

Max Activating Token Index: 696

Text #11

Max Range: 7.8363. Min Range: -7.8363

Max Act: 7.2738. Min Act: -0.1700

Data Index: 493210 (C4 (Web Text))

Max Activating Token Index: 838

Text #12

Max Range: 7.8363. Min Range: -7.8363

Max Act: 7.3050. Min Act: -0.1700

Data Index: 1307374 (C4 (Web Text))

Max Activating Token Index: 312

Text #13

Max Range: 7.8363. Min Range: -7.8363

Max Act: 7.2113. Min Act: -0.1700

Data Index: 63923 (C4 (Web Text))

Max Activating Token Index: 449

Text #14

Max Range: 7.8363. Min Range: -7.8363

Max Act: 7.1800. Min Act: -0.1700

Data Index: 1190545 (C4 (Web Text))

Max Activating Token Index: 984

Text #15

Max Range: 7.8363. Min Range: -7.8363

Max Act: 7.2113. Min Act: -0.1700

Data Index: 530475 (C4 (Web Text))

Max Activating Token Index: 287

Text #16

Max Range: 7.8363. Min Range: -7.8363

Max Act: 7.2113. Min Act: -0.1700

Data Index: 760001 (C4 (Web Text))

Max Activating Token Index: 871

Text #17

Max Range: 7.8363. Min Range: -7.8363

Max Act: 7.2113. Min Act: -0.1700

Data Index: 677442 (C4 (Web Text))

Max Activating Token Index: 946

Text #18

Max Range: 7.8363. Min Range: -7.8363

Max Act: 7.1488. Min Act: -0.1700

Data Index: 526635 (C4 (Web Text))

Max Activating Token Index: 520

Text #19

Max Range: 7.8363. Min Range: -7.8363

Max Act: 7.1488. Min Act: -0.1700

Data Index: 321016 (C4 (Web Text))

Max Activating Token Index: 327