Model: SoLU Model: 12 Layers, 6144 Neurons per Layer

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

Neuron 3821 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Hooked Transformer Loading: HookedTransformer.from_pretrained('solu-12l')

Click the toggle to see the full text!

Text #0

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.6952. Min Act: -0.0002

Data Index: 1416371 (Python Code)

Max Activating Token Index: 3

Text #1

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.6377. Min Act: -0.0002

Data Index: 1031554 (C4 (Web Text))

Max Activating Token Index: 2

Text #2

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.6377. Min Act: -0.0002

Data Index: 1379024 (Python Code)

Max Activating Token Index: 2

Text #3

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.6377. Min Act: -0.0002

Data Index: 542050 (C4 (Web Text))

Max Activating Token Index: 2

Text #4

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.6377. Min Act: -0.0002

Data Index: 283139 (C4 (Web Text))

Max Activating Token Index: 2

Text #5

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.6182. Min Act: -0.0002

Data Index: 285978 (C4 (Web Text))

Max Activating Token Index: 10

Text #6

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.5199. Min Act: -0.0002

Data Index: 37272 (C4 (Web Text))

Max Activating Token Index: 6

Text #7

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.5740. Min Act: -0.0002

Data Index: 1236895 (C4 (Web Text))

Max Activating Token Index: 823

Text #8

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.5282. Min Act: -0.0002

Data Index: 1542946 (Python Code)

Max Activating Token Index: 5

Text #9

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.3844. Min Act: -0.0002

Data Index: 1095905 (C4 (Web Text))

Max Activating Token Index: 7

Text #10

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.4436. Min Act: -0.0002

Data Index: 1030089 (C4 (Web Text))

Max Activating Token Index: 773

Text #11

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.3499. Min Act: -0.0002

Data Index: 1589728 (Python Code)

Max Activating Token Index: 3

Text #12

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.3851. Min Act: -0.0002

Data Index: 92425 (C4 (Web Text))

Max Activating Token Index: 695

Text #13

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.3196. Min Act: -0.0002

Data Index: 404935 (C4 (Web Text))

Max Activating Token Index: 661

Text #14

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.3687. Min Act: -0.0002

Data Index: 132608 (C4 (Web Text))

Max Activating Token Index: 838

Text #15

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.4097. Min Act: -0.0002

Data Index: 683859 (C4 (Web Text))

Max Activating Token Index: 5

Text #16

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.3041. Min Act: -0.0002

Data Index: 1308228 (C4 (Web Text))

Max Activating Token Index: 12

Text #17

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.2610. Min Act: -0.0002

Data Index: 1412833 (Python Code)

Max Activating Token Index: 807

Text #18

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.3089. Min Act: -0.0002

Data Index: 415969 (C4 (Web Text))

Max Activating Token Index: 778

Text #19

Max Range: 2.6952. Min Range: -2.6952

Max Act: 2.3751. Min Act: -0.0002

Data Index: 460920 (C4 (Web Text))

Max Activating Token Index: 874