Model: SoLU Model: 6 Layers, 3072 Neurons per Layer

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

Neuron 2407 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-6l')

Text #0

Max Range: 1.0784. Min Range: -1.0784

Max Act: 1.0784. Min Act: -0.0001

Data Index: 1595696 (Python Code)

Max Activating Token Index: 869

Text #1

Max Range: 1.0784. Min Range: -1.0784

Max Act: 1.0738. Min Act: -0.0001

Data Index: 1547847 (Python Code)

Max Activating Token Index: 1023

Text #2

Max Range: 1.0784. Min Range: -1.0784

Max Act: 0.9818. Min Act: -0.0001

Data Index: 1519857 (Python Code)

Max Activating Token Index: 889

Text #3

Max Range: 1.0784. Min Range: -1.0784

Max Act: 0.9330. Min Act: -0.0001

Data Index: 761245 (C4 (Web Text))

Max Activating Token Index: 1023

Text #4

Max Range: 1.0784. Min Range: -1.0784

Max Act: 0.9711. Min Act: -0.0001

Data Index: 700001 (C4 (Web Text))

Max Activating Token Index: 1023

Text #5

Max Range: 1.0784. Min Range: -1.0784

Max Act: 0.9157. Min Act: -0.0001

Data Index: 291515 (C4 (Web Text))

Max Activating Token Index: 1023

Text #6

Max Range: 1.0784. Min Range: -1.0784

Max Act: 0.9114. Min Act: -0.0001

Data Index: 1506290 (Python Code)

Max Activating Token Index: 1023

Text #7

Max Range: 1.0784. Min Range: -1.0784

Max Act: 0.8442. Min Act: -0.0001

Data Index: 488244 (C4 (Web Text))

Max Activating Token Index: 1023

Text #8

Max Range: 1.0784. Min Range: -1.0784

Max Act: 0.8947. Min Act: -0.0001

Data Index: 655016 (C4 (Web Text))

Max Activating Token Index: 1023

Text #9

Max Range: 1.0784. Min Range: -1.0784

Max Act: 0.8629. Min Act: -0.0001

Data Index: 193632 (C4 (Web Text))

Max Activating Token Index: 1023

Text #10

Max Range: 1.0784. Min Range: -1.0784

Max Act: 0.8196. Min Act: -0.0001

Data Index: 1423838 (Python Code)

Max Activating Token Index: 963

Text #11

Max Range: 1.0784. Min Range: -1.0784

Max Act: 0.7421. Min Act: -0.0001

Data Index: 1590192 (Python Code)

Max Activating Token Index: 1023

Text #12

Max Range: 1.0784. Min Range: -1.0784

Max Act: 0.6716. Min Act: -0.0001

Data Index: 828054 (C4 (Web Text))

Max Activating Token Index: 1023

Text #13

Max Range: 1.0784. Min Range: -1.0784

Max Act: 0.5846. Min Act: -0.0001

Data Index: 1064383 (C4 (Web Text))

Max Activating Token Index: 1023

Text #14

Max Range: 1.0784. Min Range: -1.0784

Max Act: 0.5856. Min Act: -0.0001

Data Index: 897951 (C4 (Web Text))

Max Activating Token Index: 1023

Text #15

Max Range: 1.0784. Min Range: -1.0784

Max Act: 0.6379. Min Act: -0.0001

Data Index: 1132322 (C4 (Web Text))

Max Activating Token Index: 1023

Text #16

Max Range: 1.0784. Min Range: -1.0784

Max Act: 0.6201. Min Act: -0.0001

Data Index: 999196 (C4 (Web Text))

Max Activating Token Index: 1023

Text #17

Max Range: 1.0784. Min Range: -1.0784

Max Act: 0.5869. Min Act: -0.0001

Data Index: 57760 (C4 (Web Text))

Max Activating Token Index: 917

Text #18

Max Range: 1.0784. Min Range: -1.0784

Max Act: 0.5086. Min Act: -0.0001

Data Index: 593827 (C4 (Web Text))

Max Activating Token Index: 1023

Text #19

Max Range: 1.0784. Min Range: -1.0784

Max Act: 0.5697. Min Act: -0.0001

Data Index: 1370514 (Python Code)

Max Activating Token Index: 1023