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

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

Neuron 3040 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: 3.3822. Min Range: -3.3822

Max Act: 3.3822. Min Act: -0.0001

Data Index: 1194449 (C4 (Web Text))

Max Activating Token Index: 847

Text #1

Max Range: 3.3822. Min Range: -3.3822

Max Act: 3.1660. Min Act: -0.0001

Data Index: 72439 (C4 (Web Text))

Max Activating Token Index: 72

Text #2

Max Range: 3.3822. Min Range: -3.3822

Max Act: 2.8935. Min Act: -0.0000

Data Index: 1247055 (C4 (Web Text))

Max Activating Token Index: 806

Text #3

Max Range: 3.3822. Min Range: -3.3822

Max Act: 2.5464. Min Act: -0.0001

Data Index: 1015932 (C4 (Web Text))

Max Activating Token Index: 674

Text #4

Max Range: 3.3822. Min Range: -3.3822

Max Act: 2.3044. Min Act: -0.0000

Data Index: 444947 (C4 (Web Text))

Max Activating Token Index: 232

Text #5

Max Range: 3.3822. Min Range: -3.3822

Max Act: 2.4124. Min Act: -0.0001

Data Index: 1182069 (C4 (Web Text))

Max Activating Token Index: 1004

Text #6

Max Range: 3.3822. Min Range: -3.3822

Max Act: 2.2535. Min Act: -0.0000

Data Index: 1251763 (C4 (Web Text))

Max Activating Token Index: 777

Text #7

Max Range: 3.3822. Min Range: -3.3822

Max Act: 2.3489. Min Act: -0.0001

Data Index: 583097 (C4 (Web Text))

Max Activating Token Index: 520

Text #8

Max Range: 3.3822. Min Range: -3.3822

Max Act: 2.1653. Min Act: -0.0001

Data Index: 146268 (C4 (Web Text))

Max Activating Token Index: 13

Text #9

Max Range: 3.3822. Min Range: -3.3822

Max Act: 2.3577. Min Act: -0.0001

Data Index: 306881 (C4 (Web Text))

Max Activating Token Index: 930

Text #10

Max Range: 3.3822. Min Range: -3.3822

Max Act: 2.3520. Min Act: -0.0001

Data Index: 276726 (C4 (Web Text))

Max Activating Token Index: 585

Text #11

Max Range: 3.3822. Min Range: -3.3822

Max Act: 2.1668. Min Act: -0.0001

Data Index: 1064626 (C4 (Web Text))

Max Activating Token Index: 75

Text #12

Max Range: 3.3822. Min Range: -3.3822

Max Act: 2.0368. Min Act: -0.0000

Data Index: 1366803 (Python Code)

Max Activating Token Index: 19

Text #13

Max Range: 3.3822. Min Range: -3.3822

Max Act: 2.0812. Min Act: -0.0001

Data Index: 1006162 (C4 (Web Text))

Max Activating Token Index: 6

Text #14

Max Range: 3.3822. Min Range: -3.3822

Max Act: 1.9871. Min Act: -0.0001

Data Index: 697004 (C4 (Web Text))

Max Activating Token Index: 172

Text #15

Max Range: 3.3822. Min Range: -3.3822

Max Act: 2.0936. Min Act: -0.0001

Data Index: 1242320 (C4 (Web Text))

Max Activating Token Index: 98

Text #16

Max Range: 3.3822. Min Range: -3.3822

Max Act: 1.9360. Min Act: -0.0000

Data Index: 749517 (C4 (Web Text))

Max Activating Token Index: 242

Text #17

Max Range: 3.3822. Min Range: -3.3822

Max Act: 1.9851. Min Act: -0.0001

Data Index: 226479 (C4 (Web Text))

Max Activating Token Index: 564

Text #18

Max Range: 3.3822. Min Range: -3.3822

Max Act: 1.8602. Min Act: -0.0001

Data Index: 904544 (C4 (Web Text))

Max Activating Token Index: 938

Text #19

Max Range: 3.3822. Min Range: -3.3822

Max Act: 2.2008. Min Act: -0.0001

Data Index: 1253056 (C4 (Web Text))

Max Activating Token Index: 989