Model: SoLU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 1962 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-3l')

Text #0

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0567. Min Act: -0.0003

Data Index: 759901 (C4 (Web Text))

Max Activating Token Index: 431

Text #1

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0567. Min Act: -0.0003

Data Index: 478006 (C4 (Web Text))

Max Activating Token Index: 716

Text #2

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0530. Min Act: -0.0003

Data Index: 1334737 (C4 (Web Text))

Max Activating Token Index: 749

Text #3

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0520. Min Act: -0.0003

Data Index: 707321 (C4 (Web Text))

Max Activating Token Index: 258

Text #4

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0536. Min Act: -0.0003

Data Index: 468645 (C4 (Web Text))

Max Activating Token Index: 301

Text #5

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0520. Min Act: -0.0004

Data Index: 87268 (C4 (Web Text))

Max Activating Token Index: 462

Text #6

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0506. Min Act: -0.0004

Data Index: 1019761 (C4 (Web Text))

Max Activating Token Index: 374

Text #7

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0520. Min Act: -0.0003

Data Index: 721627 (C4 (Web Text))

Max Activating Token Index: 321

Text #8

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0477. Min Act: -0.0004

Data Index: 1269327 (C4 (Web Text))

Max Activating Token Index: 802

Text #9

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0503. Min Act: -0.0004

Data Index: 55435 (C4 (Web Text))

Max Activating Token Index: 593

Text #10

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0485. Min Act: -0.0003

Data Index: 1230651 (C4 (Web Text))

Max Activating Token Index: 846

Text #11

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0479. Min Act: -0.0004

Data Index: 112707 (C4 (Web Text))

Max Activating Token Index: 860

Text #12

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0485. Min Act: -0.0003

Data Index: 18122 (C4 (Web Text))

Max Activating Token Index: 166

Text #13

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0477. Min Act: -0.0003

Data Index: 791150 (C4 (Web Text))

Max Activating Token Index: 459

Text #14

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0468. Min Act: -0.0003

Data Index: 311134 (C4 (Web Text))

Max Activating Token Index: 1002

Text #15

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0474. Min Act: -0.0003

Data Index: 1447260 (Python Code)

Max Activating Token Index: 604

Text #16

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0473. Min Act: -0.0003

Data Index: 987646 (C4 (Web Text))

Max Activating Token Index: 176

Text #17

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0472. Min Act: -0.0003

Data Index: 872738 (C4 (Web Text))

Max Activating Token Index: 506

Text #18

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0461. Min Act: -0.0004

Data Index: 1259593 (C4 (Web Text))

Max Activating Token Index: 661

Text #19

Max Range: 0.0567. Min Range: -0.0567

Max Act: 0.0466. Min Act: -0.0003

Data Index: 337996 (C4 (Web Text))

Max Activating Token Index: 947