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

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

Neuron 954 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: 3.1873. Min Range: -3.1873

Max Act: 3.1873. Min Act: -0.1700

Data Index: 1650164 (Python Code)

Max Activating Token Index: 934

Text #1

Max Range: 3.1873. Min Range: -3.1873

Max Act: 3.1873. Min Act: -0.1700

Data Index: 1264951 (C4 (Web Text))

Max Activating Token Index: 480

Text #2

Max Range: 3.1873. Min Range: -3.1873

Max Act: 3.1558. Min Act: -0.1700

Data Index: 954087 (C4 (Web Text))

Max Activating Token Index: 564

Text #3

Max Range: 3.1873. Min Range: -3.1873

Max Act: 3.0612. Min Act: -0.1700

Data Index: 129229 (C4 (Web Text))

Max Activating Token Index: 678

Text #4

Max Range: 3.1873. Min Range: -3.1873

Max Act: 3.0454. Min Act: -0.1700

Data Index: 590926 (C4 (Web Text))

Max Activating Token Index: 850

Text #5

Max Range: 3.1873. Min Range: -3.1873

Max Act: 3.0454. Min Act: -0.1700

Data Index: 1155076 (C4 (Web Text))

Max Activating Token Index: 614

Text #6

Max Range: 3.1873. Min Range: -3.1873

Max Act: 3.0296. Min Act: -0.1700

Data Index: 1208853 (C4 (Web Text))

Max Activating Token Index: 833

Text #7

Max Range: 3.1873. Min Range: -3.1873

Max Act: 2.9980. Min Act: -0.1700

Data Index: 539292 (C4 (Web Text))

Max Activating Token Index: 742

Text #8

Max Range: 3.1873. Min Range: -3.1873

Max Act: 3.0454. Min Act: -0.1700

Data Index: 966609 (C4 (Web Text))

Max Activating Token Index: 507

Text #9

Max Range: 3.1873. Min Range: -3.1873

Max Act: 3.0296. Min Act: -0.1700

Data Index: 1059143 (C4 (Web Text))

Max Activating Token Index: 573

Text #10

Max Range: 3.1873. Min Range: -3.1873

Max Act: 2.9980. Min Act: -0.1700

Data Index: 1241627 (C4 (Web Text))

Max Activating Token Index: 570

Text #11

Max Range: 3.1873. Min Range: -3.1873

Max Act: 2.9822. Min Act: -0.1700

Data Index: 471528 (C4 (Web Text))

Max Activating Token Index: 529

Text #12

Max Range: 3.1873. Min Range: -3.1873

Max Act: 2.9980. Min Act: -0.1700

Data Index: 1369526 (Python Code)

Max Activating Token Index: 627

Text #13

Max Range: 3.1873. Min Range: -3.1873

Max Act: 2.9822. Min Act: -0.1700

Data Index: 303068 (C4 (Web Text))

Max Activating Token Index: 744

Text #14

Max Range: 3.1873. Min Range: -3.1873

Max Act: 2.9664. Min Act: -0.1700

Data Index: 1450926 (Python Code)

Max Activating Token Index: 589

Text #15

Max Range: 3.1873. Min Range: -3.1873

Max Act: 2.9505. Min Act: -0.1700

Data Index: 1135769 (C4 (Web Text))

Max Activating Token Index: 977

Text #16

Max Range: 3.1873. Min Range: -3.1873

Max Act: 2.9505. Min Act: -0.1700

Data Index: 1405818 (Python Code)

Max Activating Token Index: 776

Text #17

Max Range: 3.1873. Min Range: -3.1873

Max Act: 2.9188. Min Act: -0.1700

Data Index: 1371132 (Python Code)

Max Activating Token Index: 663

Text #18

Max Range: 3.1873. Min Range: -3.1873

Max Act: 2.8871. Min Act: -0.1700

Data Index: 1451555 (Python Code)

Max Activating Token Index: 515

Text #19

Max Range: 3.1873. Min Range: -3.1873

Max Act: 2.9030. Min Act: -0.1700

Data Index: 195634 (C4 (Web Text))

Max Activating Token Index: 769