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

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

Neuron 65 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: 1.2964. Min Range: -1.2964

Max Act: 1.2273. Min Act: -0.0001

Data Index: 1320111 (C4 (Web Text))

Max Activating Token Index: 178

Text #1

Max Range: 1.2964. Min Range: -1.2964

Max Act: 1.2964. Min Act: -0.0001

Data Index: 1196255 (C4 (Web Text))

Max Activating Token Index: 34

Text #2

Max Range: 1.2964. Min Range: -1.2964

Max Act: 1.1433. Min Act: -0.0001

Data Index: 520432 (C4 (Web Text))

Max Activating Token Index: 293

Text #3

Max Range: 1.2964. Min Range: -1.2964

Max Act: 1.0434. Min Act: -0.0001

Data Index: 901918 (C4 (Web Text))

Max Activating Token Index: 978

Text #4

Max Range: 1.2964. Min Range: -1.2964

Max Act: 1.0216. Min Act: -0.0001

Data Index: 752824 (C4 (Web Text))

Max Activating Token Index: 61

Text #5

Max Range: 1.2964. Min Range: -1.2964

Max Act: 0.8768. Min Act: -0.0001

Data Index: 1246759 (C4 (Web Text))

Max Activating Token Index: 715

Text #6

Max Range: 1.2964. Min Range: -1.2964

Max Act: 0.9473. Min Act: -0.0001

Data Index: 658891 (C4 (Web Text))

Max Activating Token Index: 639

Text #7

Max Range: 1.2964. Min Range: -1.2964

Max Act: 0.9391. Min Act: -0.0001

Data Index: 351082 (C4 (Web Text))

Max Activating Token Index: 545

Text #8

Max Range: 1.2964. Min Range: -1.2964

Max Act: 0.9469. Min Act: -0.0001

Data Index: 545841 (C4 (Web Text))

Max Activating Token Index: 304

Text #9

Max Range: 1.2964. Min Range: -1.2964

Max Act: 0.8887. Min Act: -0.0001

Data Index: 1233513 (C4 (Web Text))

Max Activating Token Index: 458

Text #10

Max Range: 1.2964. Min Range: -1.2964

Max Act: 0.8960. Min Act: -0.0001

Data Index: 87115 (C4 (Web Text))

Max Activating Token Index: 909

Text #11

Max Range: 1.2964. Min Range: -1.2964

Max Act: 0.8164. Min Act: -0.0001

Data Index: 1397805 (Python Code)

Max Activating Token Index: 351

Text #12

Max Range: 1.2964. Min Range: -1.2964

Max Act: 0.8812. Min Act: -0.0001

Data Index: 181336 (C4 (Web Text))

Max Activating Token Index: 519

Text #13

Max Range: 1.2964. Min Range: -1.2964

Max Act: 0.8238. Min Act: -0.0001

Data Index: 322481 (C4 (Web Text))

Max Activating Token Index: 544

Text #14

Max Range: 1.2964. Min Range: -1.2964

Max Act: 0.8469. Min Act: -0.0001

Data Index: 486510 (C4 (Web Text))

Max Activating Token Index: 318

Text #15

Max Range: 1.2964. Min Range: -1.2964

Max Act: 0.8124. Min Act: -0.0001

Data Index: 14268 (C4 (Web Text))

Max Activating Token Index: 601

Text #16

Max Range: 1.2964. Min Range: -1.2964

Max Act: 0.7914. Min Act: -0.0001

Data Index: 383631 (C4 (Web Text))

Max Activating Token Index: 976

Text #17

Max Range: 1.2964. Min Range: -1.2964

Max Act: 0.7996. Min Act: -0.0001

Data Index: 1168139 (C4 (Web Text))

Max Activating Token Index: 84

Text #18

Max Range: 1.2964. Min Range: -1.2964

Max Act: 0.8192. Min Act: -0.0001

Data Index: 651327 (C4 (Web Text))

Max Activating Token Index: 404

Text #19

Max Range: 1.2964. Min Range: -1.2964

Max Act: 0.7930. Min Act: -0.0001

Data Index: 528057 (C4 (Web Text))

Max Activating Token Index: 28