Model: GELU Model: 2 Layers, 2048 Neurons per Layer

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

Neuron 1980 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-2l')

Text #0

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.6238. Min Act: -0.1700

Data Index: 1084624 (C4 (Web Text))

Max Activating Token Index: 368

Text #1

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.6395. Min Act: -0.1700

Data Index: 1611444 (Python Code)

Max Activating Token Index: 3

Text #2

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.6395. Min Act: -0.1700

Data Index: 1511125 (Python Code)

Max Activating Token Index: 3

Text #3

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.6395. Min Act: -0.1700

Data Index: 1498519 (Python Code)

Max Activating Token Index: 3

Text #4

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.6238. Min Act: -0.1700

Data Index: 297131 (C4 (Web Text))

Max Activating Token Index: 991

Text #5

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.6082. Min Act: -0.1700

Data Index: 1344770 (C4 (Web Text))

Max Activating Token Index: 147

Text #6

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.6082. Min Act: -0.1700

Data Index: 1601539 (Python Code)

Max Activating Token Index: 573

Text #7

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.5925. Min Act: -0.1700

Data Index: 30069 (C4 (Web Text))

Max Activating Token Index: 685

Text #8

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.5456. Min Act: -0.1700

Data Index: 975214 (C4 (Web Text))

Max Activating Token Index: 1012

Text #9

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.5142. Min Act: -0.1700

Data Index: 1123946 (C4 (Web Text))

Max Activating Token Index: 183

Text #10

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.5456. Min Act: -0.1700

Data Index: 1401779 (Python Code)

Max Activating Token Index: 859

Text #11

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.5142. Min Act: -0.1700

Data Index: 197415 (C4 (Web Text))

Max Activating Token Index: 702

Text #12

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.5456. Min Act: -0.1700

Data Index: 669909 (C4 (Web Text))

Max Activating Token Index: 988

Text #13

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.5299. Min Act: -0.1700

Data Index: 1593906 (Python Code)

Max Activating Token Index: 382

Text #14

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.4672. Min Act: -0.1700

Data Index: 538932 (C4 (Web Text))

Max Activating Token Index: 931

Text #15

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.4986. Min Act: -0.1700

Data Index: 1101470 (C4 (Web Text))

Max Activating Token Index: 870

Text #16

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.5142. Min Act: -0.1700

Data Index: 605456 (C4 (Web Text))

Max Activating Token Index: 189

Text #17

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.4986. Min Act: -0.1700

Data Index: 1514868 (Python Code)

Max Activating Token Index: 886

Text #18

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.4986. Min Act: -0.1700

Data Index: 1526171 (Python Code)

Max Activating Token Index: 741

Text #19

Max Range: 3.6395. Min Range: -3.6395

Max Act: 3.4829. Min Act: -0.1700

Data Index: 1449557 (Python Code)

Max Activating Token Index: 442