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

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

Neuron 1084 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: 5.1460. Min Range: -5.1460

Max Act: 5.1460. Min Act: -0.1700

Data Index: 1438498 (Python Code)

Max Activating Token Index: 383

Text #1

Max Range: 5.1460. Min Range: -5.1460

Max Act: 4.7085. Min Act: -0.1700

Data Index: 234126 (C4 (Web Text))

Max Activating Token Index: 794

Text #2

Max Range: 5.1460. Min Range: -5.1460

Max Act: 4.6460. Min Act: -0.1700

Data Index: 1546667 (Python Code)

Max Activating Token Index: 879

Text #3

Max Range: 5.1460. Min Range: -5.1460

Max Act: 4.5210. Min Act: -0.1700

Data Index: 1050583 (C4 (Web Text))

Max Activating Token Index: 288

Text #4

Max Range: 5.1460. Min Range: -5.1460

Max Act: 4.5522. Min Act: -0.1700

Data Index: 1379916 (Python Code)

Max Activating Token Index: 82

Text #5

Max Range: 5.1460. Min Range: -5.1460

Max Act: 4.5522. Min Act: -0.1700

Data Index: 103486 (C4 (Web Text))

Max Activating Token Index: 229

Text #6

Max Range: 5.1460. Min Range: -5.1460

Max Act: 4.4897. Min Act: -0.1700

Data Index: 769955 (C4 (Web Text))

Max Activating Token Index: 547

Text #7

Max Range: 5.1460. Min Range: -5.1460

Max Act: 4.4585. Min Act: -0.1700

Data Index: 796369 (C4 (Web Text))

Max Activating Token Index: 928

Text #8

Max Range: 5.1460. Min Range: -5.1460

Max Act: 4.4585. Min Act: -0.1700

Data Index: 705736 (C4 (Web Text))

Max Activating Token Index: 270

Text #9

Max Range: 5.1460. Min Range: -5.1460

Max Act: 4.4272. Min Act: -0.1700

Data Index: 61780 (C4 (Web Text))

Max Activating Token Index: 867

Text #10

Max Range: 5.1460. Min Range: -5.1460

Max Act: 4.3647. Min Act: -0.1700

Data Index: 1504389 (Python Code)

Max Activating Token Index: 363

Text #11

Max Range: 5.1460. Min Range: -5.1460

Max Act: 4.3960. Min Act: -0.1700

Data Index: 1444524 (Python Code)

Max Activating Token Index: 73

Text #12

Max Range: 5.1460. Min Range: -5.1460

Max Act: 4.3960. Min Act: -0.1700

Data Index: 1398087 (Python Code)

Max Activating Token Index: 428

Text #13

Max Range: 5.1460. Min Range: -5.1460

Max Act: 4.3335. Min Act: -0.1700

Data Index: 1552385 (Python Code)

Max Activating Token Index: 409

Text #14

Max Range: 5.1460. Min Range: -5.1460

Max Act: 4.3335. Min Act: -0.1700

Data Index: 310539 (C4 (Web Text))

Max Activating Token Index: 452

Text #15

Max Range: 5.1460. Min Range: -5.1460

Max Act: 4.3335. Min Act: -0.1700

Data Index: 445574 (C4 (Web Text))

Max Activating Token Index: 209

Text #16

Max Range: 5.1460. Min Range: -5.1460

Max Act: 4.3022. Min Act: -0.1700

Data Index: 358188 (C4 (Web Text))

Max Activating Token Index: 542

Text #17

Max Range: 5.1460. Min Range: -5.1460

Max Act: 4.2710. Min Act: -0.1700

Data Index: 1606582 (Python Code)

Max Activating Token Index: 212

Text #18

Max Range: 5.1460. Min Range: -5.1460

Max Act: 4.3022. Min Act: -0.1700

Data Index: 297707 (C4 (Web Text))

Max Activating Token Index: 598

Text #19

Max Range: 5.1460. Min Range: -5.1460

Max Act: 4.3022. Min Act: -0.1700

Data Index: 805473 (C4 (Web Text))

Max Activating Token Index: 431