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

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

Neuron 1713 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: 3.8428. Min Range: -3.8428

Max Act: 3.8428. Min Act: -0.1700

Data Index: 1514355 (Python Code)

Max Activating Token Index: 159

Text #1

Max Range: 3.8428. Min Range: -3.8428

Max Act: 3.7802. Min Act: -0.1700

Data Index: 1645257 (Python Code)

Max Activating Token Index: 94

Text #2

Max Range: 3.8428. Min Range: -3.8428

Max Act: 3.5924. Min Act: -0.1700

Data Index: 1099848 (C4 (Web Text))

Max Activating Token Index: 515

Text #3

Max Range: 3.8428. Min Range: -3.8428

Max Act: 3.6550. Min Act: -0.1700

Data Index: 1430625 (Python Code)

Max Activating Token Index: 813

Text #4

Max Range: 3.8428. Min Range: -3.8428

Max Act: 3.5141. Min Act: -0.1700

Data Index: 1229049 (C4 (Web Text))

Max Activating Token Index: 684

Text #5

Max Range: 3.8428. Min Range: -3.8428

Max Act: 3.5611. Min Act: -0.1700

Data Index: 1521203 (Python Code)

Max Activating Token Index: 487

Text #6

Max Range: 3.8428. Min Range: -3.8428

Max Act: 3.5454. Min Act: -0.1700

Data Index: 1531726 (Python Code)

Max Activating Token Index: 6

Text #7

Max Range: 3.8428. Min Range: -3.8428

Max Act: 3.5611. Min Act: -0.1700

Data Index: 1600636 (Python Code)

Max Activating Token Index: 301

Text #8

Max Range: 3.8428. Min Range: -3.8428

Max Act: 3.5611. Min Act: -0.1700

Data Index: 1564376 (Python Code)

Max Activating Token Index: 523

Text #9

Max Range: 3.8428. Min Range: -3.8428

Max Act: 3.5141. Min Act: -0.1700

Data Index: 1526198 (Python Code)

Max Activating Token Index: 44

Text #10

Max Range: 3.8428. Min Range: -3.8428

Max Act: 3.5454. Min Act: -0.1700

Data Index: 1362857 (Python Code)

Max Activating Token Index: 13

Text #11

Max Range: 3.8428. Min Range: -3.8428

Max Act: 3.4828. Min Act: -0.1700

Data Index: 1613708 (Python Code)

Max Activating Token Index: 646

Text #12

Max Range: 3.8428. Min Range: -3.8428

Max Act: 3.4984. Min Act: -0.1700

Data Index: 1601830 (Python Code)

Max Activating Token Index: 39

Text #13

Max Range: 3.8428. Min Range: -3.8428

Max Act: 3.5141. Min Act: -0.1700

Data Index: 580747 (C4 (Web Text))

Max Activating Token Index: 986

Text #14

Max Range: 3.8428. Min Range: -3.8428

Max Act: 3.4984. Min Act: -0.1700

Data Index: 522488 (C4 (Web Text))

Max Activating Token Index: 121

Text #15

Max Range: 3.8428. Min Range: -3.8428

Max Act: 3.4984. Min Act: -0.1700

Data Index: 1520021 (Python Code)

Max Activating Token Index: 598

Text #16

Max Range: 3.8428. Min Range: -3.8428

Max Act: 3.5141. Min Act: -0.1700

Data Index: 1545723 (Python Code)

Max Activating Token Index: 95

Text #17

Max Range: 3.8428. Min Range: -3.8428

Max Act: 3.5141. Min Act: -0.1700

Data Index: 1427227 (Python Code)

Max Activating Token Index: 924

Text #18

Max Range: 3.8428. Min Range: -3.8428

Max Act: 3.4514. Min Act: -0.1700

Data Index: 1391437 (Python Code)

Max Activating Token Index: 461

Text #19

Max Range: 3.8428. Min Range: -3.8428

Max Act: 3.4514. Min Act: -0.1700

Data Index: 1490619 (Python Code)

Max Activating Token Index: 958