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

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

Neuron 1200 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: 3.3649. Min Range: -3.3649

Max Act: 3.3649. Min Act: -0.1700

Data Index: 553894 (C4 (Web Text))

Max Activating Token Index: 798

Text #1

Max Range: 3.3649. Min Range: -3.3649

Max Act: 3.2235. Min Act: -0.1700

Data Index: 1223514 (C4 (Web Text))

Max Activating Token Index: 375

Text #2

Max Range: 3.3649. Min Range: -3.3649

Max Act: 2.9554. Min Act: -0.1700

Data Index: 216335 (C4 (Web Text))

Max Activating Token Index: 376

Text #3

Max Range: 3.3649. Min Range: -3.3649

Max Act: 2.9237. Min Act: -0.1700

Data Index: 754241 (C4 (Web Text))

Max Activating Token Index: 14

Text #4

Max Range: 3.3649. Min Range: -3.3649

Max Act: 2.8761. Min Act: -0.1700

Data Index: 89230 (C4 (Web Text))

Max Activating Token Index: 870

Text #5

Max Range: 3.3649. Min Range: -3.3649

Max Act: 2.8602. Min Act: -0.1700

Data Index: 266038 (C4 (Web Text))

Max Activating Token Index: 779

Text #6

Max Range: 3.3649. Min Range: -3.3649

Max Act: 2.8125. Min Act: -0.1700

Data Index: 420808 (C4 (Web Text))

Max Activating Token Index: 408

Text #7

Max Range: 3.3649. Min Range: -3.3649

Max Act: 2.8125. Min Act: -0.1700

Data Index: 696770 (C4 (Web Text))

Max Activating Token Index: 4

Text #8

Max Range: 3.3649. Min Range: -3.3649

Max Act: 2.7807. Min Act: -0.1700

Data Index: 416047 (C4 (Web Text))

Max Activating Token Index: 505

Text #9

Max Range: 3.3649. Min Range: -3.3649

Max Act: 2.7966. Min Act: -0.1700

Data Index: 1025205 (C4 (Web Text))

Max Activating Token Index: 962

Text #10

Max Range: 3.3649. Min Range: -3.3649

Max Act: 2.7807. Min Act: -0.1700

Data Index: 187517 (C4 (Web Text))

Max Activating Token Index: 209

Text #11

Max Range: 3.3649. Min Range: -3.3649

Max Act: 2.7966. Min Act: -0.1700

Data Index: 525670 (C4 (Web Text))

Max Activating Token Index: 347

Text #12

Max Range: 3.3649. Min Range: -3.3649

Max Act: 2.7966. Min Act: -0.1700

Data Index: 864673 (C4 (Web Text))

Max Activating Token Index: 13

Text #13

Max Range: 3.3649. Min Range: -3.3649

Max Act: 2.7647. Min Act: -0.1700

Data Index: 1134990 (C4 (Web Text))

Max Activating Token Index: 6

Text #14

Max Range: 3.3649. Min Range: -3.3649

Max Act: 2.7966. Min Act: -0.1700

Data Index: 43346 (C4 (Web Text))

Max Activating Token Index: 126

Text #15

Max Range: 3.3649. Min Range: -3.3649

Max Act: 2.7647. Min Act: -0.1700

Data Index: 1101617 (C4 (Web Text))

Max Activating Token Index: 269

Text #16

Max Range: 3.3649. Min Range: -3.3649

Max Act: 2.7488. Min Act: -0.1700

Data Index: 1440108 (Python Code)

Max Activating Token Index: 13

Text #17

Max Range: 3.3649. Min Range: -3.3649

Max Act: 2.7647. Min Act: -0.1700

Data Index: 963552 (C4 (Web Text))

Max Activating Token Index: 977

Text #18

Max Range: 3.3649. Min Range: -3.3649

Max Act: 2.7647. Min Act: -0.1700

Data Index: 536754 (C4 (Web Text))

Max Activating Token Index: 3

Text #19

Max Range: 3.3649. Min Range: -3.3649

Max Act: 2.7647. Min Act: -0.1700

Data Index: 226795 (C4 (Web Text))

Max Activating Token Index: 790