Model: SoLU Model: 4 Layers, 2048 Neurons per Layer

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

Neuron 1974 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-4l')

Text #0

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0549. Min Act: -0.0005

Data Index: 320031 (C4 (Web Text))

Max Activating Token Index: 366

Text #1

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0535. Min Act: -0.0005

Data Index: 1111824 (C4 (Web Text))

Max Activating Token Index: 246

Text #2

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0542. Min Act: -0.0005

Data Index: 1089981 (C4 (Web Text))

Max Activating Token Index: 990

Text #3

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0500. Min Act: -0.0005

Data Index: 1413640 (Python Code)

Max Activating Token Index: 396

Text #4

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0516. Min Act: -0.0005

Data Index: 306054 (C4 (Web Text))

Max Activating Token Index: 303

Text #5

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0490. Min Act: -0.0005

Data Index: 734304 (C4 (Web Text))

Max Activating Token Index: 329

Text #6

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0491. Min Act: -0.0005

Data Index: 1138325 (C4 (Web Text))

Max Activating Token Index: 824

Text #7

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0478. Min Act: -0.0005

Data Index: 573536 (C4 (Web Text))

Max Activating Token Index: 84

Text #8

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0487. Min Act: -0.0005

Data Index: 2820 (C4 (Web Text))

Max Activating Token Index: 118

Text #9

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0513. Min Act: -0.0005

Data Index: 1638585 (Python Code)

Max Activating Token Index: 973

Text #10

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0464. Min Act: -0.0005

Data Index: 391486 (C4 (Web Text))

Max Activating Token Index: 686

Text #11

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0480. Min Act: -0.0005

Data Index: 862026 (C4 (Web Text))

Max Activating Token Index: 75

Text #12

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0454. Min Act: -0.0005

Data Index: 1401020 (Python Code)

Max Activating Token Index: 686

Text #13

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0469. Min Act: -0.0005

Data Index: 620247 (C4 (Web Text))

Max Activating Token Index: 691

Text #14

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0482. Min Act: -0.0005

Data Index: 708224 (C4 (Web Text))

Max Activating Token Index: 787

Text #15

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0455. Min Act: -0.0005

Data Index: 1072592 (C4 (Web Text))

Max Activating Token Index: 291

Text #16

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0454. Min Act: -0.0005

Data Index: 374806 (C4 (Web Text))

Max Activating Token Index: 699

Text #17

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0471. Min Act: -0.0005

Data Index: 531132 (C4 (Web Text))

Max Activating Token Index: 88

Text #18

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0471. Min Act: -0.0005

Data Index: 538950 (C4 (Web Text))

Max Activating Token Index: 198

Text #19

Max Range: 0.0549. Min Range: -0.0549

Max Act: 0.0443. Min Act: -0.0005

Data Index: 1322073 (C4 (Web Text))

Max Activating Token Index: 1021