Model: GELU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 1633 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-3l')

Text #0

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.5107. Min Act: -0.1700

Data Index: 621146 (C4 (Web Text))

Max Activating Token Index: 56

Text #1

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.4323. Min Act: -0.1700

Data Index: 534904 (C4 (Web Text))

Max Activating Token Index: 809

Text #2

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.4323. Min Act: -0.1700

Data Index: 743887 (C4 (Web Text))

Max Activating Token Index: 417

Text #3

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.4009. Min Act: -0.1700

Data Index: 304288 (C4 (Web Text))

Max Activating Token Index: 132

Text #4

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.3696. Min Act: -0.1700

Data Index: 1554438 (Python Code)

Max Activating Token Index: 747

Text #5

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.4009. Min Act: -0.1700

Data Index: 894474 (C4 (Web Text))

Max Activating Token Index: 472

Text #6

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.3853. Min Act: -0.1700

Data Index: 1273700 (C4 (Web Text))

Max Activating Token Index: 717

Text #7

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.4009. Min Act: -0.1700

Data Index: 300958 (C4 (Web Text))

Max Activating Token Index: 594

Text #8

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.4009. Min Act: -0.1700

Data Index: 1610977 (Python Code)

Max Activating Token Index: 522

Text #9

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.3853. Min Act: -0.1700

Data Index: 1449019 (Python Code)

Max Activating Token Index: 911

Text #10

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.3696. Min Act: -0.1700

Data Index: 1653206 (Python Code)

Max Activating Token Index: 342

Text #11

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.3539. Min Act: -0.1700

Data Index: 1645964 (Python Code)

Max Activating Token Index: 696

Text #12

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.3382. Min Act: -0.1700

Data Index: 729327 (C4 (Web Text))

Max Activating Token Index: 18

Text #13

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.3696. Min Act: -0.1700

Data Index: 1335521 (C4 (Web Text))

Max Activating Token Index: 370

Text #14

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.3382. Min Act: -0.1700

Data Index: 275640 (C4 (Web Text))

Max Activating Token Index: 746

Text #15

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.3225. Min Act: -0.1700

Data Index: 948594 (C4 (Web Text))

Max Activating Token Index: 304

Text #16

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.3068. Min Act: -0.1700

Data Index: 777314 (C4 (Web Text))

Max Activating Token Index: 435

Text #17

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.3225. Min Act: -0.1700

Data Index: 1638849 (Python Code)

Max Activating Token Index: 575

Text #18

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.3225. Min Act: -0.1700

Data Index: 659242 (C4 (Web Text))

Max Activating Token Index: 448

Text #19

Max Range: 3.5107. Min Range: -3.5107

Max Act: 3.3382. Min Act: -0.1700

Data Index: 308162 (C4 (Web Text))

Max Activating Token Index: 381