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

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

Neuron 1897 in Layer 0

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-4l')

Text #0

Max Range: 3.0099. Min Range: -3.0099

Max Act: 2.9941. Min Act: -0.1700

Data Index: 207942 (C4 (Web Text))

Max Activating Token Index: 555

Text #1

Max Range: 3.0099. Min Range: -3.0099

Max Act: 3.0099. Min Act: -0.1700

Data Index: 469395 (C4 (Web Text))

Max Activating Token Index: 600

Text #2

Max Range: 3.0099. Min Range: -3.0099

Max Act: 2.9782. Min Act: -0.1700

Data Index: 881476 (C4 (Web Text))

Max Activating Token Index: 608

Text #3

Max Range: 3.0099. Min Range: -3.0099

Max Act: 2.9466. Min Act: -0.1700

Data Index: 1276377 (C4 (Web Text))

Max Activating Token Index: 528

Text #4

Max Range: 3.0099. Min Range: -3.0099

Max Act: 2.8832. Min Act: -0.1700

Data Index: 1618503 (Python Code)

Max Activating Token Index: 20

Text #5

Max Range: 3.0099. Min Range: -3.0099

Max Act: 2.8990. Min Act: -0.1700

Data Index: 1180652 (C4 (Web Text))

Max Activating Token Index: 410

Text #6

Max Range: 3.0099. Min Range: -3.0099

Max Act: 2.8832. Min Act: -0.1700

Data Index: 1210281 (C4 (Web Text))

Max Activating Token Index: 757

Text #7

Max Range: 3.0099. Min Range: -3.0099

Max Act: 2.8673. Min Act: -0.1700

Data Index: 908879 (C4 (Web Text))

Max Activating Token Index: 485

Text #8

Max Range: 3.0099. Min Range: -3.0099

Max Act: 2.8673. Min Act: -0.1700

Data Index: 118867 (C4 (Web Text))

Max Activating Token Index: 808

Text #9

Max Range: 3.0099. Min Range: -3.0099

Max Act: 2.8514. Min Act: -0.1700

Data Index: 1378507 (Python Code)

Max Activating Token Index: 365

Text #10

Max Range: 3.0099. Min Range: -3.0099

Max Act: 2.8355. Min Act: -0.1700

Data Index: 1492097 (Python Code)

Max Activating Token Index: 21

Text #11

Max Range: 3.0099. Min Range: -3.0099

Max Act: 2.8037. Min Act: -0.1700

Data Index: 1016914 (C4 (Web Text))

Max Activating Token Index: 781

Text #12

Max Range: 3.0099. Min Range: -3.0099

Max Act: 2.8196. Min Act: -0.1700

Data Index: 553592 (C4 (Web Text))

Max Activating Token Index: 686

Text #13

Max Range: 3.0099. Min Range: -3.0099

Max Act: 2.8355. Min Act: -0.1700

Data Index: 1059276 (C4 (Web Text))

Max Activating Token Index: 576

Text #14

Max Range: 3.0099. Min Range: -3.0099

Max Act: 2.8196. Min Act: -0.1700

Data Index: 949092 (C4 (Web Text))

Max Activating Token Index: 1010

Text #15

Max Range: 3.0099. Min Range: -3.0099

Max Act: 2.8037. Min Act: -0.1700

Data Index: 825212 (C4 (Web Text))

Max Activating Token Index: 161

Text #16

Max Range: 3.0099. Min Range: -3.0099

Max Act: 2.7878. Min Act: -0.1700

Data Index: 897377 (C4 (Web Text))

Max Activating Token Index: 911

Text #17

Max Range: 3.0099. Min Range: -3.0099

Max Act: 2.8037. Min Act: -0.1700

Data Index: 7381 (C4 (Web Text))

Max Activating Token Index: 259

Text #18

Max Range: 3.0099. Min Range: -3.0099

Max Act: 2.7878. Min Act: -0.1700

Data Index: 744099 (C4 (Web Text))

Max Activating Token Index: 528

Text #19

Max Range: 3.0099. Min Range: -3.0099

Max Act: 2.8037. Min Act: -0.1700

Data Index: 1492799 (Python Code)

Max Activating Token Index: 63