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

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

Neuron 2254 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 1.6320. Min Range: -1.6320

Max Act: 1.6320. Min Act: -0.0003

Data Index: 285066 (C4 (Web Text))

Max Activating Token Index: 564

Text #1

Max Range: 1.6320. Min Range: -1.6320

Max Act: 1.3146. Min Act: -0.0003

Data Index: 38521 (C4 (Web Text))

Max Activating Token Index: 52

Text #2

Max Range: 1.6320. Min Range: -1.6320

Max Act: 1.3280. Min Act: -0.0003

Data Index: 69015 (C4 (Web Text))

Max Activating Token Index: 240

Text #3

Max Range: 1.6320. Min Range: -1.6320

Max Act: 1.1883. Min Act: -0.0003

Data Index: 606158 (C4 (Web Text))

Max Activating Token Index: 264

Text #4

Max Range: 1.6320. Min Range: -1.6320

Max Act: 0.8617. Min Act: -0.0003

Data Index: 318111 (C4 (Web Text))

Max Activating Token Index: 765

Text #5

Max Range: 1.6320. Min Range: -1.6320

Max Act: 0.8335. Min Act: -0.0003

Data Index: 1162801 (C4 (Web Text))

Max Activating Token Index: 912

Text #6

Max Range: 1.6320. Min Range: -1.6320

Max Act: 0.9014. Min Act: -0.0003

Data Index: 770521 (C4 (Web Text))

Max Activating Token Index: 504

Text #7

Max Range: 1.6320. Min Range: -1.6320

Max Act: 0.9922. Min Act: -0.0003

Data Index: 477482 (C4 (Web Text))

Max Activating Token Index: 236

Text #8

Max Range: 1.6320. Min Range: -1.6320

Max Act: 0.9370. Min Act: -0.0003

Data Index: 115371 (C4 (Web Text))

Max Activating Token Index: 722

Text #9

Max Range: 1.6320. Min Range: -1.6320

Max Act: 0.8525. Min Act: -0.0003

Data Index: 327865 (C4 (Web Text))

Max Activating Token Index: 768

Text #10

Max Range: 1.6320. Min Range: -1.6320

Max Act: 1.0517. Min Act: -0.0003

Data Index: 243525 (C4 (Web Text))

Max Activating Token Index: 48

Text #11

Max Range: 1.6320. Min Range: -1.6320

Max Act: 0.9260. Min Act: -0.0003

Data Index: 356595 (C4 (Web Text))

Max Activating Token Index: 7

Text #12

Max Range: 1.6320. Min Range: -1.6320

Max Act: 0.9650. Min Act: -0.0003

Data Index: 1287209 (C4 (Web Text))

Max Activating Token Index: 434

Text #13

Max Range: 1.6320. Min Range: -1.6320

Max Act: 0.8243. Min Act: -0.0003

Data Index: 825500 (C4 (Web Text))

Max Activating Token Index: 939

Text #14

Max Range: 1.6320. Min Range: -1.6320

Max Act: 0.9228. Min Act: -0.0003

Data Index: 1150620 (C4 (Web Text))

Max Activating Token Index: 1017

Text #15

Max Range: 1.6320. Min Range: -1.6320

Max Act: 0.9042. Min Act: -0.0003

Data Index: 1314644 (C4 (Web Text))

Max Activating Token Index: 387

Text #16

Max Range: 1.6320. Min Range: -1.6320

Max Act: 0.8293. Min Act: -0.0003

Data Index: 1546958 (Python Code)

Max Activating Token Index: 138

Text #17

Max Range: 1.6320. Min Range: -1.6320

Max Act: 0.8710. Min Act: -0.0003

Data Index: 1116746 (C4 (Web Text))

Max Activating Token Index: 340

Text #18

Max Range: 1.6320. Min Range: -1.6320

Max Act: 0.8949. Min Act: -0.0003

Data Index: 159979 (C4 (Web Text))

Max Activating Token Index: 807

Text #19

Max Range: 1.6320. Min Range: -1.6320

Max Act: 0.8156. Min Act: -0.0003

Data Index: 860093 (C4 (Web Text))

Max Activating Token Index: 235