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

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

Neuron 167 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: 3.8507. Min Range: -3.8507

Max Act: 3.8507. Min Act: -0.1700

Data Index: 407325 (C4 (Web Text))

Max Activating Token Index: 855

Text #1

Max Range: 3.8507. Min Range: -3.8507

Max Act: 3.7569. Min Act: -0.1700

Data Index: 1577106 (Python Code)

Max Activating Token Index: 902

Text #2

Max Range: 3.8507. Min Range: -3.8507

Max Act: 3.7256. Min Act: -0.1700

Data Index: 387873 (C4 (Web Text))

Max Activating Token Index: 884

Text #3

Max Range: 3.8507. Min Range: -3.8507

Max Act: 3.7256. Min Act: -0.1700

Data Index: 1538620 (Python Code)

Max Activating Token Index: 1001

Text #4

Max Range: 3.8507. Min Range: -3.8507

Max Act: 3.8507. Min Act: -0.1700

Data Index: 870013 (C4 (Web Text))

Max Activating Token Index: 614

Text #5

Max Range: 3.8507. Min Range: -3.8507

Max Act: 3.7569. Min Act: -0.1700

Data Index: 483092 (C4 (Web Text))

Max Activating Token Index: 368

Text #6

Max Range: 3.8507. Min Range: -3.8507

Max Act: 3.8195. Min Act: -0.1700

Data Index: 909915 (C4 (Web Text))

Max Activating Token Index: 642

Text #7

Max Range: 3.8507. Min Range: -3.8507

Max Act: 3.8195. Min Act: -0.1700

Data Index: 241898 (C4 (Web Text))

Max Activating Token Index: 957

Text #8

Max Range: 3.8507. Min Range: -3.8507

Max Act: 3.7569. Min Act: -0.1700

Data Index: 376374 (C4 (Web Text))

Max Activating Token Index: 759

Text #9

Max Range: 3.8507. Min Range: -3.8507

Max Act: 3.7256. Min Act: -0.1700

Data Index: 999951 (C4 (Web Text))

Max Activating Token Index: 1022

Text #10

Max Range: 3.8507. Min Range: -3.8507

Max Act: 3.6630. Min Act: -0.1700

Data Index: 1423122 (Python Code)

Max Activating Token Index: 904

Text #11

Max Range: 3.8507. Min Range: -3.8507

Max Act: 3.6943. Min Act: -0.1700

Data Index: 1555573 (Python Code)

Max Activating Token Index: 970

Text #12

Max Range: 3.8507. Min Range: -3.8507

Max Act: 3.7256. Min Act: -0.1699

Data Index: 1563116 (Python Code)

Max Activating Token Index: 930

Text #13

Max Range: 3.8507. Min Range: -3.8507

Max Act: 3.7569. Min Act: -0.1700

Data Index: 1572275 (Python Code)

Max Activating Token Index: 760

Text #14

Max Range: 3.8507. Min Range: -3.8507

Max Act: 3.7569. Min Act: -0.1700

Data Index: 1448927 (Python Code)

Max Activating Token Index: 757

Text #15

Max Range: 3.8507. Min Range: -3.8507

Max Act: 3.6630. Min Act: -0.1700

Data Index: 1460225 (Python Code)

Max Activating Token Index: 930

Text #16

Max Range: 3.8507. Min Range: -3.8507

Max Act: 3.6943. Min Act: -0.1700

Data Index: 1464312 (Python Code)

Max Activating Token Index: 918

Text #17

Max Range: 3.8507. Min Range: -3.8507

Max Act: 3.7256. Min Act: -0.1700

Data Index: 1508754 (Python Code)

Max Activating Token Index: 914

Text #18

Max Range: 3.8507. Min Range: -3.8507

Max Act: 3.6630. Min Act: -0.1700

Data Index: 811948 (C4 (Web Text))

Max Activating Token Index: 392

Text #19

Max Range: 3.8507. Min Range: -3.8507

Max Act: 3.7256. Min Act: -0.1700

Data Index: 1163252 (C4 (Web Text))

Max Activating Token Index: 677