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

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

Neuron 32 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.6299. Min Act: -0.1700

Data Index: 1258664 (C4 (Web Text))

Max Activating Token Index: 650

Text #1

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.5986. Min Act: -0.1700

Data Index: 162160 (C4 (Web Text))

Max Activating Token Index: 890

Text #2

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.6299. Min Act: -0.1700

Data Index: 700062 (C4 (Web Text))

Max Activating Token Index: 786

Text #3

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.5829. Min Act: -0.1700

Data Index: 206775 (C4 (Web Text))

Max Activating Token Index: 329

Text #4

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.6455. Min Act: -0.1700

Data Index: 773181 (C4 (Web Text))

Max Activating Token Index: 484

Text #5

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.5986. Min Act: -0.1700

Data Index: 985408 (C4 (Web Text))

Max Activating Token Index: 631

Text #6

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.6612. Min Act: -0.1700

Data Index: 504909 (C4 (Web Text))

Max Activating Token Index: 612

Text #7

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.5986. Min Act: -0.1700

Data Index: 113011 (C4 (Web Text))

Max Activating Token Index: 526

Text #8

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.6142. Min Act: -0.1700

Data Index: 993407 (C4 (Web Text))

Max Activating Token Index: 756

Text #9

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.5986. Min Act: -0.1700

Data Index: 978236 (C4 (Web Text))

Max Activating Token Index: 718

Text #10

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.5672. Min Act: -0.1700

Data Index: 33038 (C4 (Web Text))

Max Activating Token Index: 500

Text #11

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.5202. Min Act: -0.1700

Data Index: 108579 (C4 (Web Text))

Max Activating Token Index: 702

Text #12

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.5516. Min Act: -0.1700

Data Index: 851432 (C4 (Web Text))

Max Activating Token Index: 573

Text #13

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.5202. Min Act: -0.1700

Data Index: 746755 (C4 (Web Text))

Max Activating Token Index: 621

Text #14

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.4732. Min Act: -0.1700

Data Index: 569362 (C4 (Web Text))

Max Activating Token Index: 955

Text #15

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.5202. Min Act: -0.1700

Data Index: 1289163 (C4 (Web Text))

Max Activating Token Index: 896

Text #16

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.5829. Min Act: -0.1700

Data Index: 962225 (C4 (Web Text))

Max Activating Token Index: 889

Text #17

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.5986. Min Act: -0.1700

Data Index: 1154967 (C4 (Web Text))

Max Activating Token Index: 768

Text #18

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.5986. Min Act: -0.1700

Data Index: 800915 (C4 (Web Text))

Max Activating Token Index: 600

Text #19

Max Range: 3.6612. Min Range: -3.6612

Max Act: 3.5516. Min Act: -0.1700

Data Index: 1045313 (C4 (Web Text))

Max Activating Token Index: 745