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

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

Neuron 2042 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.6665. Min Act: -0.1700

Data Index: 210834 (C4 (Web Text))

Max Activating Token Index: 280

Text #1

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.6352. Min Act: -0.1700

Data Index: 859828 (C4 (Web Text))

Max Activating Token Index: 629

Text #2

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.5882. Min Act: -0.1700

Data Index: 67799 (C4 (Web Text))

Max Activating Token Index: 113

Text #3

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.6039. Min Act: -0.1700

Data Index: 959694 (C4 (Web Text))

Max Activating Token Index: 303

Text #4

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.5725. Min Act: -0.1700

Data Index: 1198870 (C4 (Web Text))

Max Activating Token Index: 467

Text #5

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.5725. Min Act: -0.1700

Data Index: 301783 (C4 (Web Text))

Max Activating Token Index: 520

Text #6

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.5569. Min Act: -0.1700

Data Index: 1118853 (C4 (Web Text))

Max Activating Token Index: 543

Text #7

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.5569. Min Act: -0.1700

Data Index: 1553679 (Python Code)

Max Activating Token Index: 841

Text #8

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.5569. Min Act: -0.1700

Data Index: 1152946 (C4 (Web Text))

Max Activating Token Index: 542

Text #9

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.5256. Min Act: -0.1700

Data Index: 77319 (C4 (Web Text))

Max Activating Token Index: 172

Text #10

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.5256. Min Act: -0.1700

Data Index: 1259500 (C4 (Web Text))

Max Activating Token Index: 246

Text #11

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.5099. Min Act: -0.1700

Data Index: 1306137 (C4 (Web Text))

Max Activating Token Index: 746

Text #12

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.5256. Min Act: -0.1700

Data Index: 104844 (C4 (Web Text))

Max Activating Token Index: 557

Text #13

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.4785. Min Act: -0.1700

Data Index: 788065 (C4 (Web Text))

Max Activating Token Index: 751

Text #14

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.4785. Min Act: -0.1700

Data Index: 253166 (C4 (Web Text))

Max Activating Token Index: 198

Text #15

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.5099. Min Act: -0.1700

Data Index: 1424485 (Python Code)

Max Activating Token Index: 568

Text #16

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.4942. Min Act: -0.1700

Data Index: 1561716 (Python Code)

Max Activating Token Index: 956

Text #17

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.4785. Min Act: -0.1700

Data Index: 1361795 (Python Code)

Max Activating Token Index: 955

Text #18

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.4785. Min Act: -0.1700

Data Index: 1590620 (Python Code)

Max Activating Token Index: 913

Text #19

Max Range: 3.6665. Min Range: -3.6665

Max Act: 3.4942. Min Act: -0.1700

Data Index: 1418002 (Python Code)

Max Activating Token Index: 649