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

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

Neuron 1614 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 5.3897. Min Range: -5.3897

Max Act: 5.3897. Min Act: -0.1700

Data Index: 1429422 (Python Code)

Max Activating Token Index: 1017

Text #1

Max Range: 5.3897. Min Range: -5.3897

Max Act: 5.3272. Min Act: -0.1700

Data Index: 444565 (C4 (Web Text))

Max Activating Token Index: 914

Text #2

Max Range: 5.3897. Min Range: -5.3897

Max Act: 5.2022. Min Act: -0.1700

Data Index: 1516404 (Python Code)

Max Activating Token Index: 1013

Text #3

Max Range: 5.3897. Min Range: -5.3897

Max Act: 5.0772. Min Act: -0.1700

Data Index: 932301 (C4 (Web Text))

Max Activating Token Index: 861

Text #4

Max Range: 5.3897. Min Range: -5.3897

Max Act: 5.2960. Min Act: -0.1700

Data Index: 1455589 (Python Code)

Max Activating Token Index: 961

Text #5

Max Range: 5.3897. Min Range: -5.3897

Max Act: 5.2335. Min Act: -0.1700

Data Index: 653844 (C4 (Web Text))

Max Activating Token Index: 671

Text #6

Max Range: 5.3897. Min Range: -5.3897

Max Act: 5.0772. Min Act: -0.1700

Data Index: 1510517 (Python Code)

Max Activating Token Index: 162

Text #7

Max Range: 5.3897. Min Range: -5.3897

Max Act: 5.2335. Min Act: -0.1700

Data Index: 1306137 (C4 (Web Text))

Max Activating Token Index: 934

Text #8

Max Range: 5.3897. Min Range: -5.3897

Max Act: 5.2647. Min Act: -0.1698

Data Index: 174263 (C4 (Web Text))

Max Activating Token Index: 986

Text #9

Max Range: 5.3897. Min Range: -5.3897

Max Act: 5.1710. Min Act: -0.1700

Data Index: 1423827 (Python Code)

Max Activating Token Index: 929

Text #10

Max Range: 5.3897. Min Range: -5.3897

Max Act: 5.1710. Min Act: -0.1700

Data Index: 1404274 (Python Code)

Max Activating Token Index: 982

Text #11

Max Range: 5.3897. Min Range: -5.3897

Max Act: 5.0460. Min Act: -0.1700

Data Index: 1073454 (C4 (Web Text))

Max Activating Token Index: 558

Text #12

Max Range: 5.3897. Min Range: -5.3897

Max Act: 5.2960. Min Act: -0.1700

Data Index: 1579196 (Python Code)

Max Activating Token Index: 946

Text #13

Max Range: 5.3897. Min Range: -5.3897

Max Act: 5.1710. Min Act: -0.1700

Data Index: 1575627 (Python Code)

Max Activating Token Index: 903

Text #14

Max Range: 5.3897. Min Range: -5.3897

Max Act: 5.1710. Min Act: -0.1700

Data Index: 1483589 (Python Code)

Max Activating Token Index: 951

Text #15

Max Range: 5.3897. Min Range: -5.3897

Max Act: 5.0772. Min Act: -0.1700

Data Index: 432092 (C4 (Web Text))

Max Activating Token Index: 451

Text #16

Max Range: 5.3897. Min Range: -5.3897

Max Act: 5.1710. Min Act: -0.1700

Data Index: 767664 (C4 (Web Text))

Max Activating Token Index: 827

Text #17

Max Range: 5.3897. Min Range: -5.3897

Max Act: 5.1085. Min Act: -0.1700

Data Index: 267582 (C4 (Web Text))

Max Activating Token Index: 1010

Text #18

Max Range: 5.3897. Min Range: -5.3897

Max Act: 5.0460. Min Act: -0.1700

Data Index: 1508456 (Python Code)

Max Activating Token Index: 1018

Text #19

Max Range: 5.3897. Min Range: -5.3897

Max Act: 4.8585. Min Act: -0.1700

Data Index: 1642086 (Python Code)

Max Activating Token Index: 816