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

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

Neuron 3433 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.5088. Min Act: -0.0001

Data Index: 608990 (C4 (Web Text))

Max Activating Token Index: 648

Text #1

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.5235. Min Act: -0.0001

Data Index: 840349 (C4 (Web Text))

Max Activating Token Index: 505

Text #2

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.4728. Min Act: -0.0001

Data Index: 1239218 (C4 (Web Text))

Max Activating Token Index: 430

Text #3

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.4726. Min Act: -0.0001

Data Index: 1112746 (C4 (Web Text))

Max Activating Token Index: 507

Text #4

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.4495. Min Act: -0.0001

Data Index: 825400 (C4 (Web Text))

Max Activating Token Index: 740

Text #5

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.4656. Min Act: -0.0001

Data Index: 467861 (C4 (Web Text))

Max Activating Token Index: 522

Text #6

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.4533. Min Act: -0.0001

Data Index: 1477751 (Python Code)

Max Activating Token Index: 464

Text #7

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.4450. Min Act: -0.0001

Data Index: 1381800 (Python Code)

Max Activating Token Index: 709

Text #8

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.4554. Min Act: -0.0001

Data Index: 549299 (C4 (Web Text))

Max Activating Token Index: 647

Text #9

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.4684. Min Act: -0.0001

Data Index: 1266240 (C4 (Web Text))

Max Activating Token Index: 614

Text #10

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.4198. Min Act: -0.0001

Data Index: 508863 (C4 (Web Text))

Max Activating Token Index: 713

Text #11

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.4201. Min Act: -0.0001

Data Index: 1649390 (Python Code)

Max Activating Token Index: 905

Text #12

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.4489. Min Act: -0.0001

Data Index: 1381355 (Python Code)

Max Activating Token Index: 336

Text #13

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.4329. Min Act: -0.0001

Data Index: 1428232 (Python Code)

Max Activating Token Index: 634

Text #14

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.4898. Min Act: -0.0001

Data Index: 90599 (C4 (Web Text))

Max Activating Token Index: 872

Text #15

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.4161. Min Act: -0.0001

Data Index: 406731 (C4 (Web Text))

Max Activating Token Index: 443

Text #16

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.4444. Min Act: -0.0001

Data Index: 769787 (C4 (Web Text))

Max Activating Token Index: 477

Text #17

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.3730. Min Act: -0.0001

Data Index: 1167077 (C4 (Web Text))

Max Activating Token Index: 555

Text #18

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.4153. Min Act: -0.0001

Data Index: 340361 (C4 (Web Text))

Max Activating Token Index: 86

Text #19

Max Range: 0.5235. Min Range: -0.5235

Max Act: 0.4440. Min Act: -0.0001

Data Index: 1427133 (Python Code)

Max Activating Token Index: 644