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

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

Neuron 3917 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.0477. Min Range: -3.0477

Max Act: 2.8253. Min Act: -0.0001

Data Index: 1418663 (Python Code)

Max Activating Token Index: 604

Text #1

Max Range: 3.0477. Min Range: -3.0477

Max Act: 2.9317. Min Act: -0.0001

Data Index: 148053 (C4 (Web Text))

Max Activating Token Index: 491

Text #2

Max Range: 3.0477. Min Range: -3.0477

Max Act: 3.0477. Min Act: -0.0000

Data Index: 108791 (C4 (Web Text))

Max Activating Token Index: 338

Text #3

Max Range: 3.0477. Min Range: -3.0477

Max Act: 2.7401. Min Act: -0.0001

Data Index: 1209414 (C4 (Web Text))

Max Activating Token Index: 841

Text #4

Max Range: 3.0477. Min Range: -3.0477

Max Act: 2.8328. Min Act: -0.0000

Data Index: 1102663 (C4 (Web Text))

Max Activating Token Index: 667

Text #5

Max Range: 3.0477. Min Range: -3.0477

Max Act: 2.6419. Min Act: -0.0001

Data Index: 390155 (C4 (Web Text))

Max Activating Token Index: 266

Text #6

Max Range: 3.0477. Min Range: -3.0477

Max Act: 2.7588. Min Act: -0.0001

Data Index: 167042 (C4 (Web Text))

Max Activating Token Index: 160

Text #7

Max Range: 3.0477. Min Range: -3.0477

Max Act: 2.7333. Min Act: -0.0001

Data Index: 1640913 (Python Code)

Max Activating Token Index: 119

Text #8

Max Range: 3.0477. Min Range: -3.0477

Max Act: 2.5180. Min Act: -0.0001

Data Index: 842030 (C4 (Web Text))

Max Activating Token Index: 625

Text #9

Max Range: 3.0477. Min Range: -3.0477

Max Act: 2.7855. Min Act: -0.0001

Data Index: 541275 (C4 (Web Text))

Max Activating Token Index: 631

Text #10

Max Range: 3.0477. Min Range: -3.0477

Max Act: 2.6223. Min Act: -0.0001

Data Index: 722364 (C4 (Web Text))

Max Activating Token Index: 971

Text #11

Max Range: 3.0477. Min Range: -3.0477

Max Act: 2.6192. Min Act: -0.0001

Data Index: 372091 (C4 (Web Text))

Max Activating Token Index: 953

Text #12

Max Range: 3.0477. Min Range: -3.0477

Max Act: 2.4430. Min Act: -0.0000

Data Index: 530241 (C4 (Web Text))

Max Activating Token Index: 468

Text #13

Max Range: 3.0477. Min Range: -3.0477

Max Act: 2.7087. Min Act: -0.0001

Data Index: 192004 (C4 (Web Text))

Max Activating Token Index: 299

Text #14

Max Range: 3.0477. Min Range: -3.0477

Max Act: 2.3414. Min Act: -0.0001

Data Index: 1155164 (C4 (Web Text))

Max Activating Token Index: 24

Text #15

Max Range: 3.0477. Min Range: -3.0477

Max Act: 2.3165. Min Act: -0.0000

Data Index: 190045 (C4 (Web Text))

Max Activating Token Index: 630

Text #16

Max Range: 3.0477. Min Range: -3.0477

Max Act: 2.4373. Min Act: -0.0001

Data Index: 1276105 (C4 (Web Text))

Max Activating Token Index: 481

Text #17

Max Range: 3.0477. Min Range: -3.0477

Max Act: 2.4228. Min Act: -0.0001

Data Index: 1077615 (C4 (Web Text))

Max Activating Token Index: 400

Text #18

Max Range: 3.0477. Min Range: -3.0477

Max Act: 2.2857. Min Act: -0.0001

Data Index: 1620466 (Python Code)

Max Activating Token Index: 873

Text #19

Max Range: 3.0477. Min Range: -3.0477

Max Act: 2.3917. Min Act: -0.0001

Data Index: 1542090 (Python Code)

Max Activating Token Index: 58