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

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

Neuron 1566 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 5.3047. Min Range: -5.3047

Max Act: 5.3047. Min Act: -0.1700

Data Index: 1490780 (Python Code)

Max Activating Token Index: 267

Text #1

Max Range: 5.3047. Min Range: -5.3047

Max Act: 5.1641. Min Act: -0.1700

Data Index: 1504050 (Python Code)

Max Activating Token Index: 405

Text #2

Max Range: 5.3047. Min Range: -5.3047

Max Act: 5.1641. Min Act: -0.1700

Data Index: 1473493 (Python Code)

Max Activating Token Index: 598

Text #3

Max Range: 5.3047. Min Range: -5.3047

Max Act: 5.1797. Min Act: -0.1700

Data Index: 351042 (C4 (Web Text))

Max Activating Token Index: 458

Text #4

Max Range: 5.3047. Min Range: -5.3047

Max Act: 5.2109. Min Act: -0.1700

Data Index: 1560030 (Python Code)

Max Activating Token Index: 967

Text #5

Max Range: 5.3047. Min Range: -5.3047

Max Act: 5.1641. Min Act: -0.1700

Data Index: 1245453 (C4 (Web Text))

Max Activating Token Index: 268

Text #6

Max Range: 5.3047. Min Range: -5.3047

Max Act: 5.0391. Min Act: -0.1700

Data Index: 1523613 (Python Code)

Max Activating Token Index: 619

Text #7

Max Range: 5.3047. Min Range: -5.3047

Max Act: 5.0391. Min Act: -0.1700

Data Index: 1532637 (Python Code)

Max Activating Token Index: 677

Text #8

Max Range: 5.3047. Min Range: -5.3047

Max Act: 5.0859. Min Act: -0.1700

Data Index: 1541752 (Python Code)

Max Activating Token Index: 886

Text #9

Max Range: 5.3047. Min Range: -5.3047

Max Act: 5.0547. Min Act: -0.1700

Data Index: 1546542 (Python Code)

Max Activating Token Index: 806

Text #10

Max Range: 5.3047. Min Range: -5.3047

Max Act: 5.0234. Min Act: -0.1700

Data Index: 552536 (C4 (Web Text))

Max Activating Token Index: 380

Text #11

Max Range: 5.3047. Min Range: -5.3047

Max Act: 5.0391. Min Act: -0.1700

Data Index: 44057 (C4 (Web Text))

Max Activating Token Index: 914

Text #12

Max Range: 5.3047. Min Range: -5.3047

Max Act: 5.0234. Min Act: -0.1700

Data Index: 1387640 (Python Code)

Max Activating Token Index: 782

Text #13

Max Range: 5.3047. Min Range: -5.3047

Max Act: 5.0391. Min Act: -0.1700

Data Index: 1571544 (Python Code)

Max Activating Token Index: 447

Text #14

Max Range: 5.3047. Min Range: -5.3047

Max Act: 5.0078. Min Act: -0.1700

Data Index: 1459342 (Python Code)

Max Activating Token Index: 617

Text #15

Max Range: 5.3047. Min Range: -5.3047

Max Act: 4.9609. Min Act: -0.1700

Data Index: 1410519 (Python Code)

Max Activating Token Index: 305

Text #16

Max Range: 5.3047. Min Range: -5.3047

Max Act: 4.9766. Min Act: -0.1700

Data Index: 1481270 (Python Code)

Max Activating Token Index: 834

Text #17

Max Range: 5.3047. Min Range: -5.3047

Max Act: 5.0078. Min Act: -0.1700

Data Index: 1439363 (Python Code)

Max Activating Token Index: 246

Text #18

Max Range: 5.3047. Min Range: -5.3047

Max Act: 4.9766. Min Act: -0.1700

Data Index: 1586416 (Python Code)

Max Activating Token Index: 368

Text #19

Max Range: 5.3047. Min Range: -5.3047

Max Act: 4.9453. Min Act: -0.1700

Data Index: 1465784 (Python Code)

Max Activating Token Index: 1022