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

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

Neuron 2021 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.4510. Min Act: -0.0000

Data Index: 810348 (C4 (Web Text))

Max Activating Token Index: 90

Text #1

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.4327. Min Act: -0.0000

Data Index: 98219 (C4 (Web Text))

Max Activating Token Index: 30

Text #2

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.4442. Min Act: -0.0000

Data Index: 356829 (C4 (Web Text))

Max Activating Token Index: 531

Text #3

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.4158. Min Act: -0.0000

Data Index: 1165395 (C4 (Web Text))

Max Activating Token Index: 540

Text #4

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.4047. Min Act: -0.0000

Data Index: 1115140 (C4 (Web Text))

Max Activating Token Index: 311

Text #5

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.3873. Min Act: -0.0000

Data Index: 1174334 (C4 (Web Text))

Max Activating Token Index: 311

Text #6

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.4010. Min Act: -0.0000

Data Index: 22791 (C4 (Web Text))

Max Activating Token Index: 252

Text #7

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.3891. Min Act: -0.0000

Data Index: 232309 (C4 (Web Text))

Max Activating Token Index: 373

Text #8

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.3453. Min Act: -0.0000

Data Index: 856143 (C4 (Web Text))

Max Activating Token Index: 208

Text #9

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.3817. Min Act: -0.0000

Data Index: 1229441 (C4 (Web Text))

Max Activating Token Index: 945

Text #10

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.3550. Min Act: -0.0000

Data Index: 736314 (C4 (Web Text))

Max Activating Token Index: 44

Text #11

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.3914. Min Act: -0.0000

Data Index: 660599 (C4 (Web Text))

Max Activating Token Index: 228

Text #12

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.3274. Min Act: -0.0000

Data Index: 696702 (C4 (Web Text))

Max Activating Token Index: 846

Text #13

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.3543. Min Act: -0.0000

Data Index: 850577 (C4 (Web Text))

Max Activating Token Index: 391

Text #14

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.3439. Min Act: -0.0000

Data Index: 706190 (C4 (Web Text))

Max Activating Token Index: 729

Text #15

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.3610. Min Act: -0.0000

Data Index: 490077 (C4 (Web Text))

Max Activating Token Index: 693

Text #16

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.3593. Min Act: -0.0000

Data Index: 220952 (C4 (Web Text))

Max Activating Token Index: 15

Text #17

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.3744. Min Act: -0.0001

Data Index: 1198796 (C4 (Web Text))

Max Activating Token Index: 252

Text #18

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.3828. Min Act: -0.0000

Data Index: 134805 (C4 (Web Text))

Max Activating Token Index: 530

Text #19

Max Range: 0.4510. Min Range: -0.4510

Max Act: 0.3575. Min Act: -0.0000

Data Index: 164834 (C4 (Web Text))

Max Activating Token Index: 424