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

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

Neuron 101 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: 3.8144. Min Range: -3.8144

Max Act: 3.8144. Min Act: -0.1700

Data Index: 1570986 (Python Code)

Max Activating Token Index: 403

Text #1

Max Range: 3.8144. Min Range: -3.8144

Max Act: 3.7988. Min Act: -0.1700

Data Index: 1649478 (Python Code)

Max Activating Token Index: 558

Text #2

Max Range: 3.8144. Min Range: -3.8144

Max Act: 3.7831. Min Act: -0.1700

Data Index: 248778 (C4 (Web Text))

Max Activating Token Index: 83

Text #3

Max Range: 3.8144. Min Range: -3.8144

Max Act: 3.7831. Min Act: -0.1700

Data Index: 522943 (C4 (Web Text))

Max Activating Token Index: 31

Text #4

Max Range: 3.8144. Min Range: -3.8144

Max Act: 3.7518. Min Act: -0.1700

Data Index: 1582961 (Python Code)

Max Activating Token Index: 441

Text #5

Max Range: 3.8144. Min Range: -3.8144

Max Act: 3.7518. Min Act: -0.1700

Data Index: 1426454 (Python Code)

Max Activating Token Index: 933

Text #6

Max Range: 3.8144. Min Range: -3.8144

Max Act: 3.7362. Min Act: -0.1700

Data Index: 960216 (C4 (Web Text))

Max Activating Token Index: 749

Text #7

Max Range: 3.8144. Min Range: -3.8144

Max Act: 3.7675. Min Act: -0.1700

Data Index: 1370374 (Python Code)

Max Activating Token Index: 762

Text #8

Max Range: 3.8144. Min Range: -3.8144

Max Act: 3.7675. Min Act: -0.1700

Data Index: 1537836 (Python Code)

Max Activating Token Index: 361

Text #9

Max Range: 3.8144. Min Range: -3.8144

Max Act: 3.7049. Min Act: -0.1700

Data Index: 1406050 (Python Code)

Max Activating Token Index: 338

Text #10

Max Range: 3.8144. Min Range: -3.8144

Max Act: 3.7206. Min Act: -0.1700

Data Index: 1559633 (Python Code)

Max Activating Token Index: 529

Text #11

Max Range: 3.8144. Min Range: -3.8144

Max Act: 3.7049. Min Act: -0.1699

Data Index: 742442 (C4 (Web Text))

Max Activating Token Index: 80

Text #12

Max Range: 3.8144. Min Range: -3.8144

Max Act: 3.7206. Min Act: -0.1700

Data Index: 1550670 (Python Code)

Max Activating Token Index: 701

Text #13

Max Range: 3.8144. Min Range: -3.8144

Max Act: 3.7206. Min Act: -0.1700

Data Index: 1589448 (Python Code)

Max Activating Token Index: 496

Text #14

Max Range: 3.8144. Min Range: -3.8144

Max Act: 3.6736. Min Act: -0.1700

Data Index: 1556382 (Python Code)

Max Activating Token Index: 468

Text #15

Max Range: 3.8144. Min Range: -3.8144

Max Act: 3.6893. Min Act: -0.1700

Data Index: 1442003 (Python Code)

Max Activating Token Index: 783

Text #16

Max Range: 3.8144. Min Range: -3.8144

Max Act: 3.6893. Min Act: -0.1700

Data Index: 1372417 (Python Code)

Max Activating Token Index: 679

Text #17

Max Range: 3.8144. Min Range: -3.8144

Max Act: 3.7049. Min Act: -0.1700

Data Index: 1543425 (Python Code)

Max Activating Token Index: 920

Text #18

Max Range: 3.8144. Min Range: -3.8144

Max Act: 3.6736. Min Act: -0.1700

Data Index: 1420857 (Python Code)

Max Activating Token Index: 576

Text #19

Max Range: 3.8144. Min Range: -3.8144

Max Act: 3.6893. Min Act: -0.1700

Data Index: 1361929 (Python Code)

Max Activating Token Index: 313