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

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

Neuron 1987 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 1.4611. Min Range: -1.4611

Max Act: 1.4611. Min Act: -0.0000

Data Index: 1043315 (C4 (Web Text))

Max Activating Token Index: 871

Text #1

Max Range: 1.4611. Min Range: -1.4611

Max Act: 1.2687. Min Act: -0.0000

Data Index: 1357085 (C4 (Web Text))

Max Activating Token Index: 83

Text #2

Max Range: 1.4611. Min Range: -1.4611

Max Act: 1.3181. Min Act: -0.0000

Data Index: 1209520 (C4 (Web Text))

Max Activating Token Index: 61

Text #3

Max Range: 1.4611. Min Range: -1.4611

Max Act: 1.0518. Min Act: -0.0000

Data Index: 175996 (C4 (Web Text))

Max Activating Token Index: 84

Text #4

Max Range: 1.4611. Min Range: -1.4611

Max Act: 1.1039. Min Act: -0.0000

Data Index: 675139 (C4 (Web Text))

Max Activating Token Index: 260

Text #5

Max Range: 1.4611. Min Range: -1.4611

Max Act: 0.9282. Min Act: -0.0000

Data Index: 1459958 (Python Code)

Max Activating Token Index: 622

Text #6

Max Range: 1.4611. Min Range: -1.4611

Max Act: 0.9136. Min Act: -0.0000

Data Index: 1200280 (C4 (Web Text))

Max Activating Token Index: 570

Text #7

Max Range: 1.4611. Min Range: -1.4611

Max Act: 0.8665. Min Act: -0.0000

Data Index: 24321 (C4 (Web Text))

Max Activating Token Index: 442

Text #8

Max Range: 1.4611. Min Range: -1.4611

Max Act: 0.8967. Min Act: -0.0000

Data Index: 30044 (C4 (Web Text))

Max Activating Token Index: 83

Text #9

Max Range: 1.4611. Min Range: -1.4611

Max Act: 0.8309. Min Act: -0.0000

Data Index: 733853 (C4 (Web Text))

Max Activating Token Index: 513

Text #10

Max Range: 1.4611. Min Range: -1.4611

Max Act: 0.8276. Min Act: -0.0000

Data Index: 557373 (C4 (Web Text))

Max Activating Token Index: 139

Text #11

Max Range: 1.4611. Min Range: -1.4611

Max Act: 0.8493. Min Act: -0.0000

Data Index: 453548 (C4 (Web Text))

Max Activating Token Index: 438

Text #12

Max Range: 1.4611. Min Range: -1.4611

Max Act: 0.8324. Min Act: -0.0000

Data Index: 1599973 (Python Code)

Max Activating Token Index: 703

Text #13

Max Range: 1.4611. Min Range: -1.4611

Max Act: 0.8273. Min Act: -0.0000

Data Index: 575353 (C4 (Web Text))

Max Activating Token Index: 186

Text #14

Max Range: 1.4611. Min Range: -1.4611

Max Act: 0.8186. Min Act: -0.0000

Data Index: 1416135 (Python Code)

Max Activating Token Index: 358

Text #15

Max Range: 1.4611. Min Range: -1.4611

Max Act: 0.8028. Min Act: -0.0000

Data Index: 691176 (C4 (Web Text))

Max Activating Token Index: 256

Text #16

Max Range: 1.4611. Min Range: -1.4611

Max Act: 0.7746. Min Act: -0.0000

Data Index: 1441118 (Python Code)

Max Activating Token Index: 137

Text #17

Max Range: 1.4611. Min Range: -1.4611

Max Act: 0.7221. Min Act: -0.0000

Data Index: 538322 (C4 (Web Text))

Max Activating Token Index: 43

Text #18

Max Range: 1.4611. Min Range: -1.4611

Max Act: 0.7498. Min Act: -0.0000

Data Index: 1468550 (Python Code)

Max Activating Token Index: 969

Text #19

Max Range: 1.4611. Min Range: -1.4611

Max Act: 0.7193. Min Act: -0.0000

Data Index: 1604539 (Python Code)

Max Activating Token Index: 760