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

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

Neuron 1432 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: 5.2923. Min Range: -5.2923

Max Act: 5.2923. Min Act: -0.1700

Data Index: 70123 (C4 (Web Text))

Max Activating Token Index: 942

Text #1

Max Range: 5.2923. Min Range: -5.2923

Max Act: 5.2844. Min Act: -0.1700

Data Index: 577616 (C4 (Web Text))

Max Activating Token Index: 766

Text #2

Max Range: 5.2923. Min Range: -5.2923

Max Act: 5.1673. Min Act: -0.1700

Data Index: 1301239 (C4 (Web Text))

Max Activating Token Index: 710

Text #3

Max Range: 5.2923. Min Range: -5.2923

Max Act: 5.1907. Min Act: -0.1700

Data Index: 863129 (C4 (Web Text))

Max Activating Token Index: 1004

Text #4

Max Range: 5.2923. Min Range: -5.2923

Max Act: 5.2298. Min Act: -0.1700

Data Index: 1140212 (C4 (Web Text))

Max Activating Token Index: 677

Text #5

Max Range: 5.2923. Min Range: -5.2923

Max Act: 5.1360. Min Act: -0.1700

Data Index: 127178 (C4 (Web Text))

Max Activating Token Index: 934

Text #6

Max Range: 5.2923. Min Range: -5.2923

Max Act: 5.2219. Min Act: -0.1700

Data Index: 129020 (C4 (Web Text))

Max Activating Token Index: 453

Text #7

Max Range: 5.2923. Min Range: -5.2923

Max Act: 5.1126. Min Act: -0.1700

Data Index: 1514461 (Python Code)

Max Activating Token Index: 450

Text #8

Max Range: 5.2923. Min Range: -5.2923

Max Act: 5.0032. Min Act: -0.1700

Data Index: 627305 (C4 (Web Text))

Max Activating Token Index: 908

Text #9

Max Range: 5.2923. Min Range: -5.2923

Max Act: 4.9837. Min Act: -0.1700

Data Index: 542647 (C4 (Web Text))

Max Activating Token Index: 537

Text #10

Max Range: 5.2923. Min Range: -5.2923

Max Act: 4.9680. Min Act: -0.1700

Data Index: 581454 (C4 (Web Text))

Max Activating Token Index: 441

Text #11

Max Range: 5.2923. Min Range: -5.2923

Max Act: 4.9602. Min Act: -0.1700

Data Index: 1258083 (C4 (Web Text))

Max Activating Token Index: 773

Text #12

Max Range: 5.2923. Min Range: -5.2923

Max Act: 5.0305. Min Act: -0.1700

Data Index: 1000594 (C4 (Web Text))

Max Activating Token Index: 284

Text #13

Max Range: 5.2923. Min Range: -5.2923

Max Act: 4.8899. Min Act: -0.1700

Data Index: 1045970 (C4 (Web Text))

Max Activating Token Index: 847

Text #14

Max Range: 5.2923. Min Range: -5.2923

Max Act: 4.9016. Min Act: -0.1700

Data Index: 984641 (C4 (Web Text))

Max Activating Token Index: 299

Text #15

Max Range: 5.2923. Min Range: -5.2923

Max Act: 4.8938. Min Act: -0.1700

Data Index: 1346409 (C4 (Web Text))

Max Activating Token Index: 888

Text #16

Max Range: 5.2923. Min Range: -5.2923

Max Act: 4.8938. Min Act: -0.1700

Data Index: 1027495 (C4 (Web Text))

Max Activating Token Index: 568

Text #17

Max Range: 5.2923. Min Range: -5.2923

Max Act: 4.8587. Min Act: -0.1700

Data Index: 778390 (C4 (Web Text))

Max Activating Token Index: 508

Text #18

Max Range: 5.2923. Min Range: -5.2923

Max Act: 4.8860. Min Act: -0.1700

Data Index: 292073 (C4 (Web Text))

Max Activating Token Index: 586

Text #19

Max Range: 5.2923. Min Range: -5.2923

Max Act: 4.9446. Min Act: -0.1700

Data Index: 178751 (C4 (Web Text))

Max Activating Token Index: 345