Model: SoLU Model: 12 Layers, 6144 Neurons per Layer

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

Neuron 3898 in Layer 8

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-12l')

Text #0

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.6424. Min Act: -0.0000

Data Index: 1600644 (Python Code)

Max Activating Token Index: 434

Text #1

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.5808. Min Act: -0.0000

Data Index: 1142396 (C4 (Web Text))

Max Activating Token Index: 832

Text #2

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.4730. Min Act: -0.0000

Data Index: 1646439 (Python Code)

Max Activating Token Index: 245

Text #3

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.4947. Min Act: -0.0000

Data Index: 1103105 (C4 (Web Text))

Max Activating Token Index: 463

Text #4

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.4619. Min Act: -0.0000

Data Index: 1057527 (C4 (Web Text))

Max Activating Token Index: 168

Text #5

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.4998. Min Act: -0.0000

Data Index: 316914 (C4 (Web Text))

Max Activating Token Index: 968

Text #6

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.4697. Min Act: -0.0000

Data Index: 1569670 (Python Code)

Max Activating Token Index: 985

Text #7

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.4530. Min Act: -0.0000

Data Index: 114699 (C4 (Web Text))

Max Activating Token Index: 534

Text #8

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.4670. Min Act: -0.0000

Data Index: 869090 (C4 (Web Text))

Max Activating Token Index: 76

Text #9

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.4451. Min Act: -0.0000

Data Index: 1422385 (Python Code)

Max Activating Token Index: 640

Text #10

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.4441. Min Act: -0.0000

Data Index: 670122 (C4 (Web Text))

Max Activating Token Index: 892

Text #11

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.4536. Min Act: -0.0000

Data Index: 53040 (C4 (Web Text))

Max Activating Token Index: 86

Text #12

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.4611. Min Act: -0.0000

Data Index: 1569886 (Python Code)

Max Activating Token Index: 200

Text #13

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.4168. Min Act: -0.0000

Data Index: 18669 (C4 (Web Text))

Max Activating Token Index: 732

Text #14

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.4404. Min Act: -0.0000

Data Index: 1527939 (Python Code)

Max Activating Token Index: 482

Text #15

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.4358. Min Act: -0.0000

Data Index: 1292138 (C4 (Web Text))

Max Activating Token Index: 400

Text #16

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.4288. Min Act: -0.0000

Data Index: 1401708 (Python Code)

Max Activating Token Index: 633

Text #17

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.4249. Min Act: -0.0000

Data Index: 104516 (C4 (Web Text))

Max Activating Token Index: 779

Text #18

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.3799. Min Act: -0.0000

Data Index: 638779 (C4 (Web Text))

Max Activating Token Index: 285

Text #19

Max Range: 0.6424. Min Range: -0.6424

Max Act: 0.4208. Min Act: -0.0000

Data Index: 92485 (C4 (Web Text))

Max Activating Token Index: 46