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

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

Neuron 1971 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: 3.8089. Min Range: -3.8089

Max Act: 3.8089. Min Act: -0.1700

Data Index: 1125576 (C4 (Web Text))

Max Activating Token Index: 79

Text #1

Max Range: 3.8089. Min Range: -3.8089

Max Act: 3.7150. Min Act: -0.1700

Data Index: 98316 (C4 (Web Text))

Max Activating Token Index: 991

Text #2

Max Range: 3.8089. Min Range: -3.8089

Max Act: 3.6837. Min Act: -0.1700

Data Index: 299455 (C4 (Web Text))

Max Activating Token Index: 593

Text #3

Max Range: 3.8089. Min Range: -3.8089

Max Act: 3.7150. Min Act: -0.1700

Data Index: 543657 (C4 (Web Text))

Max Activating Token Index: 186

Text #4

Max Range: 3.8089. Min Range: -3.8089

Max Act: 3.6524. Min Act: -0.1700

Data Index: 651160 (C4 (Web Text))

Max Activating Token Index: 172

Text #5

Max Range: 3.8089. Min Range: -3.8089

Max Act: 3.6211. Min Act: -0.1700

Data Index: 398591 (C4 (Web Text))

Max Activating Token Index: 1017

Text #6

Max Range: 3.8089. Min Range: -3.8089

Max Act: 3.5898. Min Act: -0.1700

Data Index: 785139 (C4 (Web Text))

Max Activating Token Index: 1008

Text #7

Max Range: 3.8089. Min Range: -3.8089

Max Act: 3.5898. Min Act: -0.1700

Data Index: 686941 (C4 (Web Text))

Max Activating Token Index: 544

Text #8

Max Range: 3.8089. Min Range: -3.8089

Max Act: 3.5898. Min Act: -0.1700

Data Index: 25904 (C4 (Web Text))

Max Activating Token Index: 942

Text #9

Max Range: 3.8089. Min Range: -3.8089

Max Act: 3.5585. Min Act: -0.1700

Data Index: 265183 (C4 (Web Text))

Max Activating Token Index: 554

Text #10

Max Range: 3.8089. Min Range: -3.8089

Max Act: 3.6211. Min Act: -0.1700

Data Index: 665271 (C4 (Web Text))

Max Activating Token Index: 190

Text #11

Max Range: 3.8089. Min Range: -3.8089

Max Act: 3.5272. Min Act: -0.1700

Data Index: 1526929 (Python Code)

Max Activating Token Index: 745

Text #12

Max Range: 3.8089. Min Range: -3.8089

Max Act: 3.5898. Min Act: -0.1700

Data Index: 1526097 (Python Code)

Max Activating Token Index: 715

Text #13

Max Range: 3.8089. Min Range: -3.8089

Max Act: 3.4958. Min Act: -0.1700

Data Index: 1014995 (C4 (Web Text))

Max Activating Token Index: 38

Text #14

Max Range: 3.8089. Min Range: -3.8089

Max Act: 3.5585. Min Act: -0.1700

Data Index: 1560603 (Python Code)

Max Activating Token Index: 870

Text #15

Max Range: 3.8089. Min Range: -3.8089

Max Act: 3.4645. Min Act: -0.1700

Data Index: 1437986 (Python Code)

Max Activating Token Index: 741

Text #16

Max Range: 3.8089. Min Range: -3.8089

Max Act: 3.5272. Min Act: -0.1700

Data Index: 1524221 (Python Code)

Max Activating Token Index: 713

Text #17

Max Range: 3.8089. Min Range: -3.8089

Max Act: 3.5272. Min Act: -0.1700

Data Index: 1401451 (Python Code)

Max Activating Token Index: 741

Text #18

Max Range: 3.8089. Min Range: -3.8089

Max Act: 3.4958. Min Act: -0.1700

Data Index: 1601085 (Python Code)

Max Activating Token Index: 295

Text #19

Max Range: 3.8089. Min Range: -3.8089

Max Act: 3.4645. Min Act: -0.1700

Data Index: 201420 (C4 (Web Text))

Max Activating Token Index: 767