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

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

Neuron 2588 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.5062. Min Act: -0.0001

Data Index: 1118129 (C4 (Web Text))

Max Activating Token Index: 676

Text #1

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.4125. Min Act: -0.0001

Data Index: 816099 (C4 (Web Text))

Max Activating Token Index: 442

Text #2

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.3840. Min Act: -0.0001

Data Index: 442885 (C4 (Web Text))

Max Activating Token Index: 255

Text #3

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.3707. Min Act: -0.0001

Data Index: 504974 (C4 (Web Text))

Max Activating Token Index: 61

Text #4

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.3656. Min Act: -0.0001

Data Index: 1293781 (C4 (Web Text))

Max Activating Token Index: 89

Text #5

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.3398. Min Act: -0.0001

Data Index: 1142715 (C4 (Web Text))

Max Activating Token Index: 771

Text #6

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.3537. Min Act: -0.0001

Data Index: 42562 (C4 (Web Text))

Max Activating Token Index: 434

Text #7

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.3386. Min Act: -0.0001

Data Index: 75488 (C4 (Web Text))

Max Activating Token Index: 143

Text #8

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.3660. Min Act: -0.0001

Data Index: 7854 (C4 (Web Text))

Max Activating Token Index: 459

Text #9

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.3358. Min Act: -0.0001

Data Index: 1182215 (C4 (Web Text))

Max Activating Token Index: 120

Text #10

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.3377. Min Act: -0.0001

Data Index: 302792 (C4 (Web Text))

Max Activating Token Index: 150

Text #11

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.3477. Min Act: -0.0001

Data Index: 1553674 (Python Code)

Max Activating Token Index: 896

Text #12

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.3262. Min Act: -0.0001

Data Index: 1317008 (C4 (Web Text))

Max Activating Token Index: 335

Text #13

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.2928. Min Act: -0.0001

Data Index: 648346 (C4 (Web Text))

Max Activating Token Index: 534

Text #14

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.2786. Min Act: -0.0001

Data Index: 119576 (C4 (Web Text))

Max Activating Token Index: 238

Text #15

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.3122. Min Act: -0.0001

Data Index: 1048479 (C4 (Web Text))

Max Activating Token Index: 315

Text #16

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.2993. Min Act: -0.0001

Data Index: 1292689 (C4 (Web Text))

Max Activating Token Index: 687

Text #17

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.2778. Min Act: -0.0001

Data Index: 1492290 (Python Code)

Max Activating Token Index: 525

Text #18

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.3356. Min Act: -0.0001

Data Index: 1000009 (C4 (Web Text))

Max Activating Token Index: 320

Text #19

Max Range: 0.5062. Min Range: -0.5062

Max Act: 0.3131. Min Act: -0.0001

Data Index: 61585 (C4 (Web Text))

Max Activating Token Index: 56