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

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

Neuron 470 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: 3.9245. Min Range: -3.9245

Max Act: 3.9245. Min Act: -0.1700

Data Index: 1642969 (Python Code)

Max Activating Token Index: 364

Text #1

Max Range: 3.9245. Min Range: -3.9245

Max Act: 3.8933. Min Act: -0.1700

Data Index: 625842 (C4 (Web Text))

Max Activating Token Index: 1023

Text #2

Max Range: 3.9245. Min Range: -3.9245

Max Act: 3.9245. Min Act: -0.1700

Data Index: 1393401 (Python Code)

Max Activating Token Index: 536

Text #3

Max Range: 3.9245. Min Range: -3.9245

Max Act: 3.9245. Min Act: -0.1700

Data Index: 1476852 (Python Code)

Max Activating Token Index: 369

Text #4

Max Range: 3.9245. Min Range: -3.9245

Max Act: 3.8776. Min Act: -0.1700

Data Index: 1545883 (Python Code)

Max Activating Token Index: 988

Text #5

Max Range: 3.9245. Min Range: -3.9245

Max Act: 3.8463. Min Act: -0.1700

Data Index: 1485667 (Python Code)

Max Activating Token Index: 360

Text #6

Max Range: 3.9245. Min Range: -3.9245

Max Act: 3.8307. Min Act: -0.1700

Data Index: 1600943 (Python Code)

Max Activating Token Index: 1020

Text #7

Max Range: 3.9245. Min Range: -3.9245

Max Act: 3.8151. Min Act: -0.1700

Data Index: 1537246 (Python Code)

Max Activating Token Index: 761

Text #8

Max Range: 3.9245. Min Range: -3.9245

Max Act: 3.7838. Min Act: -0.1700

Data Index: 401410 (C4 (Web Text))

Max Activating Token Index: 645

Text #9

Max Range: 3.9245. Min Range: -3.9245

Max Act: 3.7994. Min Act: -0.1700

Data Index: 1623533 (Python Code)

Max Activating Token Index: 648

Text #10

Max Range: 3.9245. Min Range: -3.9245

Max Act: 3.8151. Min Act: -0.1699

Data Index: 1653672 (Python Code)

Max Activating Token Index: 485

Text #11

Max Range: 3.9245. Min Range: -3.9245

Max Act: 3.7838. Min Act: -0.1700

Data Index: 1465639 (Python Code)

Max Activating Token Index: 972

Text #12

Max Range: 3.9245. Min Range: -3.9245

Max Act: 3.7994. Min Act: -0.1700

Data Index: 1464565 (Python Code)

Max Activating Token Index: 541

Text #13

Max Range: 3.9245. Min Range: -3.9245

Max Act: 3.7838. Min Act: -0.1700

Data Index: 1415940 (Python Code)

Max Activating Token Index: 498

Text #14

Max Range: 3.9245. Min Range: -3.9245

Max Act: 3.7838. Min Act: -0.1700

Data Index: 1409573 (Python Code)

Max Activating Token Index: 955

Text #15

Max Range: 3.9245. Min Range: -3.9245

Max Act: 3.7681. Min Act: -0.1700

Data Index: 1563828 (Python Code)

Max Activating Token Index: 440

Text #16

Max Range: 3.9245. Min Range: -3.9245

Max Act: 3.8151. Min Act: -0.1700

Data Index: 1497548 (Python Code)

Max Activating Token Index: 480

Text #17

Max Range: 3.9245. Min Range: -3.9245

Max Act: 3.7994. Min Act: -0.1700

Data Index: 1497692 (Python Code)

Max Activating Token Index: 303

Text #18

Max Range: 3.9245. Min Range: -3.9245

Max Act: 3.7838. Min Act: -0.1700

Data Index: 1612862 (Python Code)

Max Activating Token Index: 551

Text #19

Max Range: 3.9245. Min Range: -3.9245

Max Act: 3.7369. Min Act: -0.1700

Data Index: 1488816 (Python Code)

Max Activating Token Index: 687