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

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

Neuron 53 in Layer 11

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 1.3625. Min Range: -1.3625

Max Act: 1.3625. Min Act: -0.0000

Data Index: 382731 (C4 (Web Text))

Max Activating Token Index: 220

Text #1

Max Range: 1.3625. Min Range: -1.3625

Max Act: 1.2358. Min Act: -0.0000

Data Index: 816582 (C4 (Web Text))

Max Activating Token Index: 749

Text #2

Max Range: 1.3625. Min Range: -1.3625

Max Act: 1.1522. Min Act: 0.0000

Data Index: 1083452 (C4 (Web Text))

Max Activating Token Index: 246

Text #3

Max Range: 1.3625. Min Range: -1.3625

Max Act: 1.1846. Min Act: 0.0001

Data Index: 490773 (C4 (Web Text))

Max Activating Token Index: 196

Text #4

Max Range: 1.3625. Min Range: -1.3625

Max Act: 1.0425. Min Act: -0.0000

Data Index: 1348989 (C4 (Web Text))

Max Activating Token Index: 31

Text #5

Max Range: 1.3625. Min Range: -1.3625

Max Act: 1.0128. Min Act: -0.0000

Data Index: 160921 (C4 (Web Text))

Max Activating Token Index: 87

Text #6

Max Range: 1.3625. Min Range: -1.3625

Max Act: 1.0559. Min Act: -0.0000

Data Index: 851321 (C4 (Web Text))

Max Activating Token Index: 49

Text #7

Max Range: 1.3625. Min Range: -1.3625

Max Act: 0.9849. Min Act: -0.0000

Data Index: 928811 (C4 (Web Text))

Max Activating Token Index: 249

Text #8

Max Range: 1.3625. Min Range: -1.3625

Max Act: 0.9760. Min Act: -0.0000

Data Index: 966954 (C4 (Web Text))

Max Activating Token Index: 641

Text #9

Max Range: 1.3625. Min Range: -1.3625

Max Act: 0.9638. Min Act: -0.0000

Data Index: 884694 (C4 (Web Text))

Max Activating Token Index: 746

Text #10

Max Range: 1.3625. Min Range: -1.3625

Max Act: 0.9171. Min Act: -0.0000

Data Index: 719221 (C4 (Web Text))

Max Activating Token Index: 520

Text #11

Max Range: 1.3625. Min Range: -1.3625

Max Act: 1.0198. Min Act: -0.0000

Data Index: 1604241 (Python Code)

Max Activating Token Index: 1007

Text #12

Max Range: 1.3625. Min Range: -1.3625

Max Act: 0.9610. Min Act: -0.0000

Data Index: 296943 (C4 (Web Text))

Max Activating Token Index: 507

Text #13

Max Range: 1.3625. Min Range: -1.3625

Max Act: 0.9595. Min Act: -0.0000

Data Index: 779699 (C4 (Web Text))

Max Activating Token Index: 927

Text #14

Max Range: 1.3625. Min Range: -1.3625

Max Act: 0.9780. Min Act: 0.0000

Data Index: 407444 (C4 (Web Text))

Max Activating Token Index: 522

Text #15

Max Range: 1.3625. Min Range: -1.3625

Max Act: 0.9311. Min Act: -0.0000

Data Index: 31244 (C4 (Web Text))

Max Activating Token Index: 813

Text #16

Max Range: 1.3625. Min Range: -1.3625

Max Act: 0.8636. Min Act: -0.0000

Data Index: 1371857 (Python Code)

Max Activating Token Index: 353

Text #17

Max Range: 1.3625. Min Range: -1.3625

Max Act: 0.9305. Min Act: -0.0000

Data Index: 330036 (C4 (Web Text))

Max Activating Token Index: 181

Text #18

Max Range: 1.3625. Min Range: -1.3625

Max Act: 0.9160. Min Act: 0.0000

Data Index: 796645 (C4 (Web Text))

Max Activating Token Index: 331

Text #19

Max Range: 1.3625. Min Range: -1.3625

Max Act: 0.9109. Min Act: -0.0000

Data Index: 990361 (C4 (Web Text))

Max Activating Token Index: 371