Model: SoLU Model: 10 Layers, 5120 Neurons per Layer

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

Neuron 4897 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 3.0498. Min Range: -3.0498

Max Act: 2.9572. Min Act: -0.0001

Data Index: 67944 (C4 (Web Text))

Max Activating Token Index: 866

Text #1

Max Range: 3.0498. Min Range: -3.0498

Max Act: 2.9706. Min Act: -0.0001

Data Index: 616462 (C4 (Web Text))

Max Activating Token Index: 873

Text #2

Max Range: 3.0498. Min Range: -3.0498

Max Act: 3.0498. Min Act: -0.0001

Data Index: 1051888 (C4 (Web Text))

Max Activating Token Index: 566

Text #3

Max Range: 3.0498. Min Range: -3.0498

Max Act: 2.8396. Min Act: -0.0000

Data Index: 246621 (C4 (Web Text))

Max Activating Token Index: 10

Text #4

Max Range: 3.0498. Min Range: -3.0498

Max Act: 2.7766. Min Act: -0.0001

Data Index: 487731 (C4 (Web Text))

Max Activating Token Index: 62

Text #5

Max Range: 3.0498. Min Range: -3.0498

Max Act: 2.7488. Min Act: -0.0000

Data Index: 119481 (C4 (Web Text))

Max Activating Token Index: 10

Text #6

Max Range: 3.0498. Min Range: -3.0498

Max Act: 2.5691. Min Act: -0.0000

Data Index: 25351 (C4 (Web Text))

Max Activating Token Index: 129

Text #7

Max Range: 3.0498. Min Range: -3.0498

Max Act: 2.6431. Min Act: -0.0001

Data Index: 995670 (C4 (Web Text))

Max Activating Token Index: 65

Text #8

Max Range: 3.0498. Min Range: -3.0498

Max Act: 2.4274. Min Act: -0.0000

Data Index: 1286710 (C4 (Web Text))

Max Activating Token Index: 612

Text #9

Max Range: 3.0498. Min Range: -3.0498

Max Act: 2.5934. Min Act: -0.0001

Data Index: 819664 (C4 (Web Text))

Max Activating Token Index: 929

Text #10

Max Range: 3.0498. Min Range: -3.0498

Max Act: 2.6535. Min Act: -0.0001

Data Index: 998782 (C4 (Web Text))

Max Activating Token Index: 32

Text #11

Max Range: 3.0498. Min Range: -3.0498

Max Act: 2.7727. Min Act: -0.0001

Data Index: 819867 (C4 (Web Text))

Max Activating Token Index: 833

Text #12

Max Range: 3.0498. Min Range: -3.0498

Max Act: 2.6091. Min Act: -0.0001

Data Index: 118512 (C4 (Web Text))

Max Activating Token Index: 529

Text #13

Max Range: 3.0498. Min Range: -3.0498

Max Act: 2.4575. Min Act: -0.0001

Data Index: 107031 (C4 (Web Text))

Max Activating Token Index: 785

Text #14

Max Range: 3.0498. Min Range: -3.0498

Max Act: 2.3035. Min Act: -0.0000

Data Index: 1145421 (C4 (Web Text))

Max Activating Token Index: 452

Text #15

Max Range: 3.0498. Min Range: -3.0498

Max Act: 2.5503. Min Act: -0.0001

Data Index: 1453575 (Python Code)

Max Activating Token Index: 301

Text #16

Max Range: 3.0498. Min Range: -3.0498

Max Act: 2.3752. Min Act: -0.0001

Data Index: 328827 (C4 (Web Text))

Max Activating Token Index: 262

Text #17

Max Range: 3.0498. Min Range: -3.0498

Max Act: 2.2515. Min Act: -0.0001

Data Index: 684103 (C4 (Web Text))

Max Activating Token Index: 356

Text #18

Max Range: 3.0498. Min Range: -3.0498

Max Act: 2.2706. Min Act: -0.0000

Data Index: 562435 (C4 (Web Text))

Max Activating Token Index: 952

Text #19

Max Range: 3.0498. Min Range: -3.0498

Max Act: 2.0239. Min Act: -0.0001

Data Index: 169893 (C4 (Web Text))

Max Activating Token Index: 210