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

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

Neuron 1791 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: 3.5296. Min Range: -3.5296

Max Act: 3.5139. Min Act: -0.1700

Data Index: 748594 (C4 (Web Text))

Max Activating Token Index: 354

Text #1

Max Range: 3.5296. Min Range: -3.5296

Max Act: 3.5139. Min Act: -0.1700

Data Index: 409174 (C4 (Web Text))

Max Activating Token Index: 898

Text #2

Max Range: 3.5296. Min Range: -3.5296

Max Act: 3.5296. Min Act: -0.1700

Data Index: 1438131 (Python Code)

Max Activating Token Index: 322

Text #3

Max Range: 3.5296. Min Range: -3.5296

Max Act: 3.4982. Min Act: -0.1700

Data Index: 27082 (C4 (Web Text))

Max Activating Token Index: 764

Text #4

Max Range: 3.5296. Min Range: -3.5296

Max Act: 3.4982. Min Act: -0.1700

Data Index: 1137509 (C4 (Web Text))

Max Activating Token Index: 815

Text #5

Max Range: 3.5296. Min Range: -3.5296

Max Act: 3.4825. Min Act: -0.1700

Data Index: 130538 (C4 (Web Text))

Max Activating Token Index: 678

Text #6

Max Range: 3.5296. Min Range: -3.5296

Max Act: 3.4825. Min Act: -0.1700

Data Index: 1402824 (Python Code)

Max Activating Token Index: 257

Text #7

Max Range: 3.5296. Min Range: -3.5296

Max Act: 3.4982. Min Act: -0.1699

Data Index: 562980 (C4 (Web Text))

Max Activating Token Index: 583

Text #8

Max Range: 3.5296. Min Range: -3.5296

Max Act: 3.4669. Min Act: -0.1700

Data Index: 467342 (C4 (Web Text))

Max Activating Token Index: 335

Text #9

Max Range: 3.5296. Min Range: -3.5296

Max Act: 3.5139. Min Act: -0.1700

Data Index: 1305217 (C4 (Web Text))

Max Activating Token Index: 471

Text #10

Max Range: 3.5296. Min Range: -3.5296

Max Act: 3.4982. Min Act: -0.1700

Data Index: 213493 (C4 (Web Text))

Max Activating Token Index: 264

Text #11

Max Range: 3.5296. Min Range: -3.5296

Max Act: 3.4825. Min Act: -0.1700

Data Index: 1535516 (Python Code)

Max Activating Token Index: 262

Text #12

Max Range: 3.5296. Min Range: -3.5296

Max Act: 3.4355. Min Act: -0.1700

Data Index: 1395008 (Python Code)

Max Activating Token Index: 214

Text #13

Max Range: 3.5296. Min Range: -3.5296

Max Act: 3.4512. Min Act: -0.1700

Data Index: 1506106 (Python Code)

Max Activating Token Index: 384

Text #14

Max Range: 3.5296. Min Range: -3.5296

Max Act: 3.4669. Min Act: -0.1700

Data Index: 1632051 (Python Code)

Max Activating Token Index: 242

Text #15

Max Range: 3.5296. Min Range: -3.5296

Max Act: 3.4669. Min Act: -0.1700

Data Index: 1053227 (C4 (Web Text))

Max Activating Token Index: 466

Text #16

Max Range: 3.5296. Min Range: -3.5296

Max Act: 3.4512. Min Act: -0.1700

Data Index: 1009569 (C4 (Web Text))

Max Activating Token Index: 410

Text #17

Max Range: 3.5296. Min Range: -3.5296

Max Act: 3.4198. Min Act: -0.1700

Data Index: 386918 (C4 (Web Text))

Max Activating Token Index: 615

Text #18

Max Range: 3.5296. Min Range: -3.5296

Max Act: 3.4669. Min Act: -0.1700

Data Index: 1404030 (Python Code)

Max Activating Token Index: 78

Text #19

Max Range: 3.5296. Min Range: -3.5296

Max Act: 3.4669. Min Act: -0.1700

Data Index: 1218555 (C4 (Web Text))

Max Activating Token Index: 932