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

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

Neuron 302 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.6602. Min Act: -0.0001

Data Index: 968648 (C4 (Web Text))

Max Activating Token Index: 821

Text #1

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.6326. Min Act: -0.0002

Data Index: 946930 (C4 (Web Text))

Max Activating Token Index: 896

Text #2

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.6258. Min Act: -0.0002

Data Index: 759796 (C4 (Web Text))

Max Activating Token Index: 871

Text #3

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.6472. Min Act: -0.0001

Data Index: 124944 (C4 (Web Text))

Max Activating Token Index: 132

Text #4

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.6214. Min Act: -0.0002

Data Index: 478367 (C4 (Web Text))

Max Activating Token Index: 354

Text #5

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.6376. Min Act: -0.0001

Data Index: 983477 (C4 (Web Text))

Max Activating Token Index: 958

Text #6

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.5966. Min Act: -0.0001

Data Index: 801288 (C4 (Web Text))

Max Activating Token Index: 3

Text #7

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.6261. Min Act: -0.0002

Data Index: 1319725 (C4 (Web Text))

Max Activating Token Index: 527

Text #8

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.5747. Min Act: -0.0002

Data Index: 505264 (C4 (Web Text))

Max Activating Token Index: 349

Text #9

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.6235. Min Act: -0.0002

Data Index: 454621 (C4 (Web Text))

Max Activating Token Index: 342

Text #10

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.5603. Min Act: -0.0001

Data Index: 212578 (C4 (Web Text))

Max Activating Token Index: 647

Text #11

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.5785. Min Act: -0.0001

Data Index: 765474 (C4 (Web Text))

Max Activating Token Index: 625

Text #12

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.5739. Min Act: -0.0002

Data Index: 99388 (C4 (Web Text))

Max Activating Token Index: 421

Text #13

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.6223. Min Act: -0.0002

Data Index: 187040 (C4 (Web Text))

Max Activating Token Index: 505

Text #14

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.5741. Min Act: -0.0002

Data Index: 790957 (C4 (Web Text))

Max Activating Token Index: 768

Text #15

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.5622. Min Act: -0.0001

Data Index: 774432 (C4 (Web Text))

Max Activating Token Index: 457

Text #16

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.5568. Min Act: -0.0002

Data Index: 504949 (C4 (Web Text))

Max Activating Token Index: 584

Text #17

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.5495. Min Act: -0.0001

Data Index: 1224908 (C4 (Web Text))

Max Activating Token Index: 49

Text #18

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.5753. Min Act: -0.0001

Data Index: 1562360 (Python Code)

Max Activating Token Index: 729

Text #19

Max Range: 0.6602. Min Range: -0.6602

Max Act: 0.5368. Min Act: -0.0001

Data Index: 1075972 (C4 (Web Text))

Max Activating Token Index: 971