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

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

Neuron 4312 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.8985. Min Act: -0.0001

Data Index: 285630 (C4 (Web Text))

Max Activating Token Index: 488

Text #1

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.9340. Min Act: -0.0001

Data Index: 923415 (C4 (Web Text))

Max Activating Token Index: 722

Text #2

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.8027. Min Act: -0.0001

Data Index: 305281 (C4 (Web Text))

Max Activating Token Index: 120

Text #3

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.7888. Min Act: -0.0001

Data Index: 1375518 (Python Code)

Max Activating Token Index: 569

Text #4

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.7495. Min Act: -0.0001

Data Index: 1472300 (Python Code)

Max Activating Token Index: 821

Text #5

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.7747. Min Act: -0.0001

Data Index: 60567 (C4 (Web Text))

Max Activating Token Index: 10

Text #6

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.7873. Min Act: -0.0001

Data Index: 78519 (C4 (Web Text))

Max Activating Token Index: 590

Text #7

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.7254. Min Act: -0.0001

Data Index: 1183495 (C4 (Web Text))

Max Activating Token Index: 90

Text #8

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.7097. Min Act: -0.0001

Data Index: 245534 (C4 (Web Text))

Max Activating Token Index: 613

Text #9

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.6543. Min Act: -0.0001

Data Index: 1437207 (Python Code)

Max Activating Token Index: 175

Text #10

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.7167. Min Act: -0.0001

Data Index: 1510716 (Python Code)

Max Activating Token Index: 722

Text #11

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.7162. Min Act: -0.0001

Data Index: 1070106 (C4 (Web Text))

Max Activating Token Index: 61

Text #12

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.6514. Min Act: -0.0001

Data Index: 1602349 (Python Code)

Max Activating Token Index: 647

Text #13

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.6864. Min Act: -0.0001

Data Index: 380990 (C4 (Web Text))

Max Activating Token Index: 811

Text #14

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.7240. Min Act: -0.0001

Data Index: 1276038 (C4 (Web Text))

Max Activating Token Index: 718

Text #15

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.7247. Min Act: -0.0001

Data Index: 784094 (C4 (Web Text))

Max Activating Token Index: 322

Text #16

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.7389. Min Act: -0.0001

Data Index: 372110 (C4 (Web Text))

Max Activating Token Index: 77

Text #17

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.7186. Min Act: -0.0001

Data Index: 202949 (C4 (Web Text))

Max Activating Token Index: 385

Text #18

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.7440. Min Act: -0.0001

Data Index: 1250426 (C4 (Web Text))

Max Activating Token Index: 762

Text #19

Max Range: 0.9340. Min Range: -0.9340

Max Act: 0.7212. Min Act: -0.0001

Data Index: 744782 (C4 (Web Text))

Max Activating Token Index: 817