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

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

Neuron 1310 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: 8.3294. Min Range: -8.3294

Max Act: 8.3294. Min Act: -0.1700

Data Index: 1224191 (C4 (Web Text))

Max Activating Token Index: 144

Text #1

Max Range: 8.3294. Min Range: -8.3294

Max Act: 8.2044. Min Act: -0.1700

Data Index: 1404867 (Python Code)

Max Activating Token Index: 206

Text #2

Max Range: 8.3294. Min Range: -8.3294

Max Act: 8.2044. Min Act: -0.1700

Data Index: 1051507 (C4 (Web Text))

Max Activating Token Index: 646

Text #3

Max Range: 8.3294. Min Range: -8.3294

Max Act: 7.8294. Min Act: -0.1700

Data Index: 402031 (C4 (Web Text))

Max Activating Token Index: 415

Text #4

Max Range: 8.3294. Min Range: -8.3294

Max Act: 7.8919. Min Act: -0.1700

Data Index: 932469 (C4 (Web Text))

Max Activating Token Index: 421

Text #5

Max Range: 8.3294. Min Range: -8.3294

Max Act: 7.8294. Min Act: -0.1700

Data Index: 1386894 (Python Code)

Max Activating Token Index: 147

Text #6

Max Range: 8.3294. Min Range: -8.3294

Max Act: 7.4544. Min Act: -0.1700

Data Index: 75115 (C4 (Web Text))

Max Activating Token Index: 522

Text #7

Max Range: 8.3294. Min Range: -8.3294

Max Act: 7.3919. Min Act: -0.1700

Data Index: 571606 (C4 (Web Text))

Max Activating Token Index: 101

Text #8

Max Range: 8.3294. Min Range: -8.3294

Max Act: 7.3294. Min Act: -0.1700

Data Index: 444811 (C4 (Web Text))

Max Activating Token Index: 624

Text #9

Max Range: 8.3294. Min Range: -8.3294

Max Act: 7.2669. Min Act: -0.1700

Data Index: 1312440 (C4 (Web Text))

Max Activating Token Index: 660

Text #10

Max Range: 8.3294. Min Range: -8.3294

Max Act: 7.2669. Min Act: -0.1700

Data Index: 898375 (C4 (Web Text))

Max Activating Token Index: 462

Text #11

Max Range: 8.3294. Min Range: -8.3294

Max Act: 7.2044. Min Act: -0.1700

Data Index: 579416 (C4 (Web Text))

Max Activating Token Index: 771

Text #12

Max Range: 8.3294. Min Range: -8.3294

Max Act: 7.2044. Min Act: -0.1700

Data Index: 626387 (C4 (Web Text))

Max Activating Token Index: 381

Text #13

Max Range: 8.3294. Min Range: -8.3294

Max Act: 7.2044. Min Act: -0.1700

Data Index: 1474388 (Python Code)

Max Activating Token Index: 309

Text #14

Max Range: 8.3294. Min Range: -8.3294

Max Act: 7.1419. Min Act: -0.1700

Data Index: 1599769 (Python Code)

Max Activating Token Index: 401

Text #15

Max Range: 8.3294. Min Range: -8.3294

Max Act: 7.2669. Min Act: -0.1700

Data Index: 34340 (C4 (Web Text))

Max Activating Token Index: 210

Text #16

Max Range: 8.3294. Min Range: -8.3294

Max Act: 7.1419. Min Act: -0.1700

Data Index: 516426 (C4 (Web Text))

Max Activating Token Index: 259

Text #17

Max Range: 8.3294. Min Range: -8.3294

Max Act: 7.0794. Min Act: -0.1700

Data Index: 610833 (C4 (Web Text))

Max Activating Token Index: 382

Text #18

Max Range: 8.3294. Min Range: -8.3294

Max Act: 7.1419. Min Act: -0.1700

Data Index: 176945 (C4 (Web Text))

Max Activating Token Index: 1017

Text #19

Max Range: 8.3294. Min Range: -8.3294

Max Act: 6.9544. Min Act: -0.1700

Data Index: 319673 (C4 (Web Text))

Max Activating Token Index: 366