Model: GELU Model: 2 Layers, 2048 Neurons per Layer

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

Neuron 50 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-2l')

Text #0

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.4173. Min Act: -0.1700

Data Index: 639986 (C4 (Web Text))

Max Activating Token Index: 460

Text #1

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.4173. Min Act: -0.1700

Data Index: 958680 (C4 (Web Text))

Max Activating Token Index: 429

Text #2

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.3231. Min Act: -0.1700

Data Index: 341622 (C4 (Web Text))

Max Activating Token Index: 652

Text #3

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.3074. Min Act: -0.1700

Data Index: 162973 (C4 (Web Text))

Max Activating Token Index: 247

Text #4

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.2603. Min Act: -0.1700

Data Index: 1629361 (Python Code)

Max Activating Token Index: 675

Text #5

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.2446. Min Act: -0.1699

Data Index: 1283768 (C4 (Web Text))

Max Activating Token Index: 350

Text #6

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.2446. Min Act: -0.1700

Data Index: 559499 (C4 (Web Text))

Max Activating Token Index: 669

Text #7

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.2760. Min Act: -0.1699

Data Index: 136617 (C4 (Web Text))

Max Activating Token Index: 372

Text #8

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.2131. Min Act: -0.1700

Data Index: 865766 (C4 (Web Text))

Max Activating Token Index: 238

Text #9

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.2446. Min Act: -0.1700

Data Index: 364423 (C4 (Web Text))

Max Activating Token Index: 712

Text #10

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.2288. Min Act: -0.1700

Data Index: 182126 (C4 (Web Text))

Max Activating Token Index: 213

Text #11

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.2131. Min Act: -0.1700

Data Index: 1218126 (C4 (Web Text))

Max Activating Token Index: 461

Text #12

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.2288. Min Act: -0.1700

Data Index: 738290 (C4 (Web Text))

Max Activating Token Index: 188

Text #13

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.2288. Min Act: -0.1700

Data Index: 1504106 (Python Code)

Max Activating Token Index: 665

Text #14

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.2603. Min Act: -0.1700

Data Index: 1497544 (Python Code)

Max Activating Token Index: 915

Text #15

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.1974. Min Act: -0.1700

Data Index: 1578726 (Python Code)

Max Activating Token Index: 447

Text #16

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.2131. Min Act: -0.1700

Data Index: 615899 (C4 (Web Text))

Max Activating Token Index: 664

Text #17

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.2131. Min Act: -0.1700

Data Index: 356829 (C4 (Web Text))

Max Activating Token Index: 528

Text #18

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.1974. Min Act: -0.1700

Data Index: 1280283 (C4 (Web Text))

Max Activating Token Index: 517

Text #19

Max Range: 3.4173. Min Range: -3.4173

Max Act: 3.1974. Min Act: -0.1700

Data Index: 1235951 (C4 (Web Text))

Max Activating Token Index: 260