Model: SoLU Model: 8 Layers, 4096 Neurons per Layer

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

Neuron 3600 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.7250. Min Act: -0.0001

Data Index: 1411729 (Python Code)

Max Activating Token Index: 632

Text #1

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.6308. Min Act: -0.0001

Data Index: 1650911 (Python Code)

Max Activating Token Index: 688

Text #2

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.6923. Min Act: -0.0001

Data Index: 1636039 (Python Code)

Max Activating Token Index: 537

Text #3

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.6176. Min Act: -0.0001

Data Index: 1418052 (Python Code)

Max Activating Token Index: 81

Text #4

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.6034. Min Act: -0.0001

Data Index: 1360479 (Python Code)

Max Activating Token Index: 521

Text #5

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.5993. Min Act: -0.0001

Data Index: 1496870 (Python Code)

Max Activating Token Index: 882

Text #6

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.5964. Min Act: -0.0001

Data Index: 1522870 (Python Code)

Max Activating Token Index: 244

Text #7

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.5867. Min Act: -0.0001

Data Index: 1368171 (Python Code)

Max Activating Token Index: 892

Text #8

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.5754. Min Act: -0.0001

Data Index: 1559543 (Python Code)

Max Activating Token Index: 938

Text #9

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.5806. Min Act: -0.0001

Data Index: 1617244 (Python Code)

Max Activating Token Index: 924

Text #10

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.5376. Min Act: -0.0001

Data Index: 1499996 (Python Code)

Max Activating Token Index: 680

Text #11

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.5802. Min Act: -0.0001

Data Index: 1624639 (Python Code)

Max Activating Token Index: 484

Text #12

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.5768. Min Act: -0.0001

Data Index: 1631871 (Python Code)

Max Activating Token Index: 657

Text #13

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.5831. Min Act: -0.0001

Data Index: 1604511 (Python Code)

Max Activating Token Index: 959

Text #14

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.5611. Min Act: -0.0001

Data Index: 1368571 (Python Code)

Max Activating Token Index: 765

Text #15

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.5586. Min Act: -0.0001

Data Index: 1461501 (Python Code)

Max Activating Token Index: 916

Text #16

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.5173. Min Act: -0.0001

Data Index: 1586061 (Python Code)

Max Activating Token Index: 892

Text #17

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.5471. Min Act: -0.0001

Data Index: 1521689 (Python Code)

Max Activating Token Index: 101

Text #18

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.5420. Min Act: -0.0001

Data Index: 1456089 (Python Code)

Max Activating Token Index: 308

Text #19

Max Range: 0.7250. Min Range: -0.7250

Max Act: 0.5342. Min Act: -0.0001

Data Index: 1601111 (Python Code)

Max Activating Token Index: 744