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

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

Neuron 890 in Layer 0

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 10.3578. Min Range: -10.3578

Max Act: 10.3578. Min Act: -0.1700

Data Index: 34057 (C4 (Web Text))

Max Activating Token Index: 29

Text #1

Max Range: 10.3578. Min Range: -10.3578

Max Act: 9.7953. Min Act: -0.1700

Data Index: 1652188 (Python Code)

Max Activating Token Index: 453

Text #2

Max Range: 10.3578. Min Range: -10.3578

Max Act: 9.2328. Min Act: -0.1700

Data Index: 1483328 (Python Code)

Max Activating Token Index: 817

Text #3

Max Range: 10.3578. Min Range: -10.3578

Max Act: 9.2328. Min Act: -0.1697

Data Index: 784054 (C4 (Web Text))

Max Activating Token Index: 778

Text #4

Max Range: 10.3578. Min Range: -10.3578

Max Act: 9.1703. Min Act: -0.1700

Data Index: 1492295 (Python Code)

Max Activating Token Index: 851

Text #5

Max Range: 10.3578. Min Range: -10.3578

Max Act: 9.1078. Min Act: -0.1700

Data Index: 1423316 (Python Code)

Max Activating Token Index: 76

Text #6

Max Range: 10.3578. Min Range: -10.3578

Max Act: 8.9828. Min Act: -0.1700

Data Index: 1643779 (Python Code)

Max Activating Token Index: 675

Text #7

Max Range: 10.3578. Min Range: -10.3578

Max Act: 8.4203. Min Act: -0.1700

Data Index: 935352 (C4 (Web Text))

Max Activating Token Index: 818

Text #8

Max Range: 10.3578. Min Range: -10.3578

Max Act: 8.1078. Min Act: -0.1700

Data Index: 169061 (C4 (Web Text))

Max Activating Token Index: 967

Text #9

Max Range: 10.3578. Min Range: -10.3578

Max Act: 7.9203. Min Act: -0.1700

Data Index: 483708 (C4 (Web Text))

Max Activating Token Index: 258

Text #10

Max Range: 10.3578. Min Range: -10.3578

Max Act: 7.9203. Min Act: -0.1700

Data Index: 1213837 (C4 (Web Text))

Max Activating Token Index: 992

Text #11

Max Range: 10.3578. Min Range: -10.3578

Max Act: 7.7328. Min Act: -0.1700

Data Index: 955345 (C4 (Web Text))

Max Activating Token Index: 204

Text #12

Max Range: 10.3578. Min Range: -10.3578

Max Act: 7.5140. Min Act: -0.1700

Data Index: 1044588 (C4 (Web Text))

Max Activating Token Index: 127

Text #13

Max Range: 10.3578. Min Range: -10.3578

Max Act: 7.4203. Min Act: -0.1700

Data Index: 710261 (C4 (Web Text))

Max Activating Token Index: 735

Text #14

Max Range: 10.3578. Min Range: -10.3578

Max Act: 7.3578. Min Act: -0.1700

Data Index: 1439139 (Python Code)

Max Activating Token Index: 986

Text #15

Max Range: 10.3578. Min Range: -10.3578

Max Act: 7.2640. Min Act: -0.1700

Data Index: 1641948 (Python Code)

Max Activating Token Index: 396

Text #16

Max Range: 10.3578. Min Range: -10.3578

Max Act: 7.2640. Min Act: -0.1700

Data Index: 1601652 (Python Code)

Max Activating Token Index: 102

Text #17

Max Range: 10.3578. Min Range: -10.3578

Max Act: 7.0140. Min Act: -0.1700

Data Index: 1649474 (Python Code)

Max Activating Token Index: 810

Text #18

Max Range: 10.3578. Min Range: -10.3578

Max Act: 6.8890. Min Act: -0.1700

Data Index: 1506777 (Python Code)

Max Activating Token Index: 537

Text #19

Max Range: 10.3578. Min Range: -10.3578

Max Act: 6.8890. Min Act: -0.1700

Data Index: 1655744 (Python Code)

Max Activating Token Index: 332