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

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

Neuron 990 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: 5.7090. Min Range: -5.7090

Max Act: 5.7090. Min Act: -0.1700

Data Index: 477722 (C4 (Web Text))

Max Activating Token Index: 368

Text #1

Max Range: 5.7090. Min Range: -5.7090

Max Act: 5.5840. Min Act: -0.1700

Data Index: 831209 (C4 (Web Text))

Max Activating Token Index: 88

Text #2

Max Range: 5.7090. Min Range: -5.7090

Max Act: 5.5840. Min Act: -0.1700

Data Index: 329774 (C4 (Web Text))

Max Activating Token Index: 655

Text #3

Max Range: 5.7090. Min Range: -5.7090

Max Act: 5.5528. Min Act: -0.1700

Data Index: 277815 (C4 (Web Text))

Max Activating Token Index: 497

Text #4

Max Range: 5.7090. Min Range: -5.7090

Max Act: 5.5215. Min Act: -0.1700

Data Index: 1172303 (C4 (Web Text))

Max Activating Token Index: 847

Text #5

Max Range: 5.7090. Min Range: -5.7090

Max Act: 5.5215. Min Act: -0.1700

Data Index: 734214 (C4 (Web Text))

Max Activating Token Index: 1002

Text #6

Max Range: 5.7090. Min Range: -5.7090

Max Act: 5.4278. Min Act: -0.1700

Data Index: 663762 (C4 (Web Text))

Max Activating Token Index: 92

Text #7

Max Range: 5.7090. Min Range: -5.7090

Max Act: 5.4278. Min Act: -0.1700

Data Index: 964844 (C4 (Web Text))

Max Activating Token Index: 100

Text #8

Max Range: 5.7090. Min Range: -5.7090

Max Act: 5.3653. Min Act: -0.1700

Data Index: 919559 (C4 (Web Text))

Max Activating Token Index: 945

Text #9

Max Range: 5.7090. Min Range: -5.7090

Max Act: 5.3965. Min Act: -0.1700

Data Index: 714470 (C4 (Web Text))

Max Activating Token Index: 361

Text #10

Max Range: 5.7090. Min Range: -5.7090

Max Act: 5.3340. Min Act: -0.1700

Data Index: 459374 (C4 (Web Text))

Max Activating Token Index: 285

Text #11

Max Range: 5.7090. Min Range: -5.7090

Max Act: 5.3340. Min Act: -0.1700

Data Index: 522141 (C4 (Web Text))

Max Activating Token Index: 672

Text #12

Max Range: 5.7090. Min Range: -5.7090

Max Act: 5.1465. Min Act: -0.1700

Data Index: 1518908 (Python Code)

Max Activating Token Index: 899

Text #13

Max Range: 5.7090. Min Range: -5.7090

Max Act: 5.3028. Min Act: -0.1700

Data Index: 600656 (C4 (Web Text))

Max Activating Token Index: 586

Text #14

Max Range: 5.7090. Min Range: -5.7090

Max Act: 5.2715. Min Act: -0.1700

Data Index: 275055 (C4 (Web Text))

Max Activating Token Index: 442

Text #15

Max Range: 5.7090. Min Range: -5.7090

Max Act: 5.2403. Min Act: -0.1700

Data Index: 1163585 (C4 (Web Text))

Max Activating Token Index: 845

Text #16

Max Range: 5.7090. Min Range: -5.7090

Max Act: 5.2403. Min Act: -0.1700

Data Index: 20047 (C4 (Web Text))

Max Activating Token Index: 752

Text #17

Max Range: 5.7090. Min Range: -5.7090

Max Act: 5.2403. Min Act: -0.1700

Data Index: 841247 (C4 (Web Text))

Max Activating Token Index: 467

Text #18

Max Range: 5.7090. Min Range: -5.7090

Max Act: 5.2403. Min Act: -0.1700

Data Index: 753151 (C4 (Web Text))

Max Activating Token Index: 408

Text #19

Max Range: 5.7090. Min Range: -5.7090

Max Act: 5.3028. Min Act: -0.1700

Data Index: 403618 (C4 (Web Text))

Max Activating Token Index: 570