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

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

Neuron 3870 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 2.0065. Min Range: -2.0065

Max Act: 2.0065. Min Act: -0.0000

Data Index: 496192 (C4 (Web Text))

Max Activating Token Index: 156

Text #1

Max Range: 2.0065. Min Range: -2.0065

Max Act: 1.9594. Min Act: -0.0000

Data Index: 76493 (C4 (Web Text))

Max Activating Token Index: 185

Text #2

Max Range: 2.0065. Min Range: -2.0065

Max Act: 1.7080. Min Act: -0.0000

Data Index: 193907 (C4 (Web Text))

Max Activating Token Index: 959

Text #3

Max Range: 2.0065. Min Range: -2.0065

Max Act: 1.7073. Min Act: -0.0000

Data Index: 226292 (C4 (Web Text))

Max Activating Token Index: 716

Text #4

Max Range: 2.0065. Min Range: -2.0065

Max Act: 1.4819. Min Act: -0.0000

Data Index: 143587 (C4 (Web Text))

Max Activating Token Index: 674

Text #5

Max Range: 2.0065. Min Range: -2.0065

Max Act: 1.3422. Min Act: -0.0000

Data Index: 31237 (C4 (Web Text))

Max Activating Token Index: 807

Text #6

Max Range: 2.0065. Min Range: -2.0065

Max Act: 1.3526. Min Act: -0.0000

Data Index: 1224410 (C4 (Web Text))

Max Activating Token Index: 600

Text #7

Max Range: 2.0065. Min Range: -2.0065

Max Act: 1.2600. Min Act: -0.0000

Data Index: 778942 (C4 (Web Text))

Max Activating Token Index: 80

Text #8

Max Range: 2.0065. Min Range: -2.0065

Max Act: 1.2608. Min Act: -0.0000

Data Index: 58350 (C4 (Web Text))

Max Activating Token Index: 556

Text #9

Max Range: 2.0065. Min Range: -2.0065

Max Act: 1.3260. Min Act: -0.0000

Data Index: 1023646 (C4 (Web Text))

Max Activating Token Index: 886

Text #10

Max Range: 2.0065. Min Range: -2.0065

Max Act: 1.3217. Min Act: -0.0000

Data Index: 46267 (C4 (Web Text))

Max Activating Token Index: 179

Text #11

Max Range: 2.0065. Min Range: -2.0065

Max Act: 1.2393. Min Act: -0.0000

Data Index: 1188099 (C4 (Web Text))

Max Activating Token Index: 581

Text #12

Max Range: 2.0065. Min Range: -2.0065

Max Act: 1.2371. Min Act: -0.0000

Data Index: 392105 (C4 (Web Text))

Max Activating Token Index: 399

Text #13

Max Range: 2.0065. Min Range: -2.0065

Max Act: 1.2799. Min Act: -0.0000

Data Index: 1361945 (Python Code)

Max Activating Token Index: 978

Text #14

Max Range: 2.0065. Min Range: -2.0065

Max Act: 1.2887. Min Act: -0.0000

Data Index: 639075 (C4 (Web Text))

Max Activating Token Index: 850

Text #15

Max Range: 2.0065. Min Range: -2.0065

Max Act: 1.1782. Min Act: -0.0000

Data Index: 1529541 (Python Code)

Max Activating Token Index: 780

Text #16

Max Range: 2.0065. Min Range: -2.0065

Max Act: 1.1947. Min Act: -0.0000

Data Index: 475409 (C4 (Web Text))

Max Activating Token Index: 506

Text #17

Max Range: 2.0065. Min Range: -2.0065

Max Act: 1.2604. Min Act: -0.0000

Data Index: 1386966 (Python Code)

Max Activating Token Index: 550

Text #18

Max Range: 2.0065. Min Range: -2.0065

Max Act: 1.2679. Min Act: -0.0000

Data Index: 1400236 (Python Code)

Max Activating Token Index: 255

Text #19

Max Range: 2.0065. Min Range: -2.0065

Max Act: 1.2490. Min Act: -0.0000

Data Index: 46513 (C4 (Web Text))

Max Activating Token Index: 463