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

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

Neuron 1036 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.9578. Min Act: -0.1700

Data Index: 1537342 (Python Code)

Max Activating Token Index: 6

Text #1

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.8327. Min Act: -0.1700

Data Index: 1655770 (Python Code)

Max Activating Token Index: 595

Text #2

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.8327. Min Act: -0.1700

Data Index: 1432698 (Python Code)

Max Activating Token Index: 441

Text #3

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.8640. Min Act: -0.1700

Data Index: 1576534 (Python Code)

Max Activating Token Index: 494

Text #4

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.8640. Min Act: -0.1700

Data Index: 1423899 (Python Code)

Max Activating Token Index: 207

Text #5

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.8014. Min Act: -0.1700

Data Index: 1511920 (Python Code)

Max Activating Token Index: 80

Text #6

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.8014. Min Act: -0.1700

Data Index: 1378821 (Python Code)

Max Activating Token Index: 953

Text #7

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.7075. Min Act: -0.1700

Data Index: 1476252 (Python Code)

Max Activating Token Index: 7

Text #8

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.7075. Min Act: -0.1700

Data Index: 1570279 (Python Code)

Max Activating Token Index: 706

Text #9

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.7075. Min Act: -0.1700

Data Index: 1654802 (Python Code)

Max Activating Token Index: 353

Text #10

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.6762. Min Act: -0.1700

Data Index: 1546522 (Python Code)

Max Activating Token Index: 348

Text #11

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.6762. Min Act: -0.1700

Data Index: 1411140 (Python Code)

Max Activating Token Index: 233

Text #12

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.6136. Min Act: -0.1700

Data Index: 1638773 (Python Code)

Max Activating Token Index: 445

Text #13

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.6449. Min Act: -0.1700

Data Index: 1563720 (Python Code)

Max Activating Token Index: 674

Text #14

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.6136. Min Act: -0.1700

Data Index: 1439879 (Python Code)

Max Activating Token Index: 382

Text #15

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.6136. Min Act: -0.1700

Data Index: 1530153 (Python Code)

Max Activating Token Index: 35

Text #16

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.6136. Min Act: -0.1700

Data Index: 1454954 (Python Code)

Max Activating Token Index: 410

Text #17

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.5823. Min Act: -0.1700

Data Index: 1560309 (Python Code)

Max Activating Token Index: 482

Text #18

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.6762. Min Act: -0.1700

Data Index: 1629988 (Python Code)

Max Activating Token Index: 167

Text #19

Max Range: 3.9578. Min Range: -3.9578

Max Act: 3.5823. Min Act: -0.1700

Data Index: 1595132 (Python Code)

Max Activating Token Index: 737