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

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

Neuron 717 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.3614. Min Act: -0.1700

Data Index: 924414 (C4 (Web Text))

Max Activating Token Index: 274

Text #1

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.3301. Min Act: -0.1700

Data Index: 1614194 (Python Code)

Max Activating Token Index: 259

Text #2

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.2676. Min Act: -0.1700

Data Index: 182742 (C4 (Web Text))

Max Activating Token Index: 356

Text #3

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.2364. Min Act: -0.1700

Data Index: 1615386 (Python Code)

Max Activating Token Index: 252

Text #4

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.2364. Min Act: -0.1700

Data Index: 1523859 (Python Code)

Max Activating Token Index: 361

Text #5

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.2676. Min Act: -0.1700

Data Index: 1413077 (Python Code)

Max Activating Token Index: 845

Text #6

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.2051. Min Act: -0.1700

Data Index: 1574087 (Python Code)

Max Activating Token Index: 419

Text #7

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.1739. Min Act: -0.1700

Data Index: 695377 (C4 (Web Text))

Max Activating Token Index: 726

Text #8

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.1114. Min Act: -0.1700

Data Index: 1406552 (Python Code)

Max Activating Token Index: 267

Text #9

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.1114. Min Act: -0.1700

Data Index: 1647373 (Python Code)

Max Activating Token Index: 923

Text #10

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.1114. Min Act: -0.1700

Data Index: 627549 (C4 (Web Text))

Max Activating Token Index: 297

Text #11

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.0489. Min Act: -0.1700

Data Index: 575559 (C4 (Web Text))

Max Activating Token Index: 733

Text #12

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.0489. Min Act: -0.1700

Data Index: 1627244 (Python Code)

Max Activating Token Index: 418

Text #13

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.0801. Min Act: -0.1700

Data Index: 1466350 (Python Code)

Max Activating Token Index: 263

Text #14

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.0801. Min Act: -0.1700

Data Index: 1643053 (Python Code)

Max Activating Token Index: 363

Text #15

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.0489. Min Act: -0.1700

Data Index: 1170851 (C4 (Web Text))

Max Activating Token Index: 948

Text #16

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.0489. Min Act: -0.1700

Data Index: 1458020 (Python Code)

Max Activating Token Index: 863

Text #17

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.0176. Min Act: -0.1700

Data Index: 1421701 (Python Code)

Max Activating Token Index: 912

Text #18

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.0489. Min Act: -0.1700

Data Index: 1567026 (Python Code)

Max Activating Token Index: 938

Text #19

Max Range: 5.3614. Min Range: -5.3614

Max Act: 5.0489. Min Act: -0.1700

Data Index: 1565056 (Python Code)

Max Activating Token Index: 272