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

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

Neuron 1718 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: 3.6358. Min Range: -3.6358

Max Act: 3.6358. Min Act: -0.1700

Data Index: 1504957 (Python Code)

Max Activating Token Index: 346

Text #1

Max Range: 3.6358. Min Range: -3.6358

Max Act: 3.5731. Min Act: -0.1700

Data Index: 1533306 (Python Code)

Max Activating Token Index: 513

Text #2

Max Range: 3.6358. Min Range: -3.6358

Max Act: 3.5731. Min Act: -0.1700

Data Index: 1485220 (Python Code)

Max Activating Token Index: 591

Text #3

Max Range: 3.6358. Min Range: -3.6358

Max Act: 3.4478. Min Act: -0.1700

Data Index: 23040 (C4 (Web Text))

Max Activating Token Index: 758

Text #4

Max Range: 3.6358. Min Range: -3.6358

Max Act: 3.4478. Min Act: -0.1700

Data Index: 1053245 (C4 (Web Text))

Max Activating Token Index: 610

Text #5

Max Range: 3.6358. Min Range: -3.6358

Max Act: 3.4791. Min Act: -0.1700

Data Index: 1415453 (Python Code)

Max Activating Token Index: 127

Text #6

Max Range: 3.6358. Min Range: -3.6358

Max Act: 3.4478. Min Act: -0.1700

Data Index: 1488401 (Python Code)

Max Activating Token Index: 199

Text #7

Max Range: 3.6358. Min Range: -3.6358

Max Act: 3.4164. Min Act: -0.1700

Data Index: 149715 (C4 (Web Text))

Max Activating Token Index: 360

Text #8

Max Range: 3.6358. Min Range: -3.6358

Max Act: 3.3537. Min Act: -0.1700

Data Index: 1363126 (Python Code)

Max Activating Token Index: 129

Text #9

Max Range: 3.6358. Min Range: -3.6358

Max Act: 3.3537. Min Act: -0.1700

Data Index: 887577 (C4 (Web Text))

Max Activating Token Index: 436

Text #10

Max Range: 3.6358. Min Range: -3.6358

Max Act: 3.3851. Min Act: -0.1700

Data Index: 594173 (C4 (Web Text))

Max Activating Token Index: 709

Text #11

Max Range: 3.6358. Min Range: -3.6358

Max Act: 3.3851. Min Act: -0.1700

Data Index: 1199162 (C4 (Web Text))

Max Activating Token Index: 587

Text #12

Max Range: 3.6358. Min Range: -3.6358

Max Act: 3.3851. Min Act: -0.1700

Data Index: 460741 (C4 (Web Text))

Max Activating Token Index: 775

Text #13

Max Range: 3.6358. Min Range: -3.6358

Max Act: 3.3537. Min Act: -0.1700

Data Index: 933237 (C4 (Web Text))

Max Activating Token Index: 161

Text #14

Max Range: 3.6358. Min Range: -3.6358

Max Act: 3.3223. Min Act: -0.1700

Data Index: 1420401 (Python Code)

Max Activating Token Index: 432

Text #15

Max Range: 3.6358. Min Range: -3.6358

Max Act: 3.3537. Min Act: -0.1700

Data Index: 435051 (C4 (Web Text))

Max Activating Token Index: 767

Text #16

Max Range: 3.6358. Min Range: -3.6358

Max Act: 3.3537. Min Act: -0.1700

Data Index: 229339 (C4 (Web Text))

Max Activating Token Index: 131

Text #17

Max Range: 3.6358. Min Range: -3.6358

Max Act: 3.3223. Min Act: -0.1700

Data Index: 1069678 (C4 (Web Text))

Max Activating Token Index: 667

Text #18

Max Range: 3.6358. Min Range: -3.6358

Max Act: 3.3223. Min Act: -0.1700

Data Index: 103019 (C4 (Web Text))

Max Activating Token Index: 601

Text #19

Max Range: 3.6358. Min Range: -3.6358

Max Act: 3.3537. Min Act: -0.1700

Data Index: 105026 (C4 (Web Text))

Max Activating Token Index: 56