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

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

Neuron 1943 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: 0.8347. Min Range: -0.8347

Max Act: 0.8347. Min Act: -0.0001

Data Index: 695379 (C4 (Web Text))

Max Activating Token Index: 674

Text #1

Max Range: 0.8347. Min Range: -0.8347

Max Act: 0.7093. Min Act: -0.0001

Data Index: 389181 (C4 (Web Text))

Max Activating Token Index: 631

Text #2

Max Range: 0.8347. Min Range: -0.8347

Max Act: 0.7567. Min Act: -0.0001

Data Index: 440718 (C4 (Web Text))

Max Activating Token Index: 785

Text #3

Max Range: 0.8347. Min Range: -0.8347

Max Act: 0.6884. Min Act: -0.0001

Data Index: 144755 (C4 (Web Text))

Max Activating Token Index: 708

Text #4

Max Range: 0.8347. Min Range: -0.8347

Max Act: 0.6184. Min Act: -0.0001

Data Index: 988635 (C4 (Web Text))

Max Activating Token Index: 967

Text #5

Max Range: 0.8347. Min Range: -0.8347

Max Act: 0.5886. Min Act: -0.0001

Data Index: 1604189 (Python Code)

Max Activating Token Index: 960

Text #6

Max Range: 0.8347. Min Range: -0.8347

Max Act: 0.6457. Min Act: -0.0001

Data Index: 816120 (C4 (Web Text))

Max Activating Token Index: 613

Text #7

Max Range: 0.8347. Min Range: -0.8347

Max Act: 0.6302. Min Act: -0.0001

Data Index: 1278898 (C4 (Web Text))

Max Activating Token Index: 865

Text #8

Max Range: 0.8347. Min Range: -0.8347

Max Act: 0.5841. Min Act: -0.0001

Data Index: 53403 (C4 (Web Text))

Max Activating Token Index: 903

Text #9

Max Range: 0.8347. Min Range: -0.8347

Max Act: 0.5786. Min Act: -0.0001

Data Index: 598141 (C4 (Web Text))

Max Activating Token Index: 286

Text #10

Max Range: 0.8347. Min Range: -0.8347

Max Act: 0.5410. Min Act: -0.0001

Data Index: 1421027 (Python Code)

Max Activating Token Index: 981

Text #11

Max Range: 0.8347. Min Range: -0.8347

Max Act: 0.5700. Min Act: -0.0001

Data Index: 543278 (C4 (Web Text))

Max Activating Token Index: 38

Text #12

Max Range: 0.8347. Min Range: -0.8347

Max Act: 0.5074. Min Act: -0.0001

Data Index: 273641 (C4 (Web Text))

Max Activating Token Index: 958

Text #13

Max Range: 0.8347. Min Range: -0.8347

Max Act: 0.5596. Min Act: -0.0001

Data Index: 1047405 (C4 (Web Text))

Max Activating Token Index: 970

Text #14

Max Range: 0.8347. Min Range: -0.8347

Max Act: 0.4713. Min Act: -0.0001

Data Index: 207181 (C4 (Web Text))

Max Activating Token Index: 596

Text #15

Max Range: 0.8347. Min Range: -0.8347

Max Act: 0.5124. Min Act: -0.0001

Data Index: 1205561 (C4 (Web Text))

Max Activating Token Index: 505

Text #16

Max Range: 0.8347. Min Range: -0.8347

Max Act: 0.5266. Min Act: -0.0001

Data Index: 51332 (C4 (Web Text))

Max Activating Token Index: 1021

Text #17

Max Range: 0.8347. Min Range: -0.8347

Max Act: 0.5363. Min Act: -0.0001

Data Index: 941844 (C4 (Web Text))

Max Activating Token Index: 965

Text #18

Max Range: 0.8347. Min Range: -0.8347

Max Act: 0.5214. Min Act: -0.0001

Data Index: 1459389 (Python Code)

Max Activating Token Index: 412

Text #19

Max Range: 0.8347. Min Range: -0.8347

Max Act: 0.5561. Min Act: -0.0001

Data Index: 82716 (C4 (Web Text))

Max Activating Token Index: 95