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

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

Neuron 2539 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: 0.9531. Min Range: -0.9531

Max Act: 0.9531. Min Act: -0.0001

Data Index: 270836 (C4 (Web Text))

Max Activating Token Index: 378

Text #1

Max Range: 0.9531. Min Range: -0.9531

Max Act: 0.9176. Min Act: -0.0001

Data Index: 904264 (C4 (Web Text))

Max Activating Token Index: 378

Text #2

Max Range: 0.9531. Min Range: -0.9531

Max Act: 0.9289. Min Act: -0.0001

Data Index: 555139 (C4 (Web Text))

Max Activating Token Index: 1023

Text #3

Max Range: 0.9531. Min Range: -0.9531

Max Act: 0.9213. Min Act: -0.0001

Data Index: 155606 (C4 (Web Text))

Max Activating Token Index: 998

Text #4

Max Range: 0.9531. Min Range: -0.9531

Max Act: 0.8990. Min Act: -0.0001

Data Index: 1257515 (C4 (Web Text))

Max Activating Token Index: 439

Text #5

Max Range: 0.9531. Min Range: -0.9531

Max Act: 0.9058. Min Act: -0.0001

Data Index: 840160 (C4 (Web Text))

Max Activating Token Index: 382

Text #6

Max Range: 0.9531. Min Range: -0.9531

Max Act: 0.7830. Min Act: -0.0001

Data Index: 1170200 (C4 (Web Text))

Max Activating Token Index: 622

Text #7

Max Range: 0.9531. Min Range: -0.9531

Max Act: 0.8290. Min Act: -0.0001

Data Index: 694685 (C4 (Web Text))

Max Activating Token Index: 495

Text #8

Max Range: 0.9531. Min Range: -0.9531

Max Act: 0.7935. Min Act: -0.0001

Data Index: 680647 (C4 (Web Text))

Max Activating Token Index: 994

Text #9

Max Range: 0.9531. Min Range: -0.9531

Max Act: 0.8021. Min Act: -0.0001

Data Index: 1490210 (Python Code)

Max Activating Token Index: 409

Text #10

Max Range: 0.9531. Min Range: -0.9531

Max Act: 0.8057. Min Act: -0.0001

Data Index: 27707 (C4 (Web Text))

Max Activating Token Index: 19

Text #11

Max Range: 0.9531. Min Range: -0.9531

Max Act: 0.7420. Min Act: -0.0001

Data Index: 471704 (C4 (Web Text))

Max Activating Token Index: 896

Text #12

Max Range: 0.9531. Min Range: -0.9531

Max Act: 0.7536. Min Act: -0.0001

Data Index: 310435 (C4 (Web Text))

Max Activating Token Index: 614

Text #13

Max Range: 0.9531. Min Range: -0.9531

Max Act: 0.7414. Min Act: -0.0001

Data Index: 572767 (C4 (Web Text))

Max Activating Token Index: 903

Text #14

Max Range: 0.9531. Min Range: -0.9531

Max Act: 0.7421. Min Act: -0.0001

Data Index: 1493550 (Python Code)

Max Activating Token Index: 1023

Text #15

Max Range: 0.9531. Min Range: -0.9531

Max Act: 0.7427. Min Act: -0.0001

Data Index: 1085194 (C4 (Web Text))

Max Activating Token Index: 523

Text #16

Max Range: 0.9531. Min Range: -0.9531

Max Act: 0.7521. Min Act: -0.0001

Data Index: 286328 (C4 (Web Text))

Max Activating Token Index: 61

Text #17

Max Range: 0.9531. Min Range: -0.9531

Max Act: 0.7549. Min Act: -0.0001

Data Index: 711128 (C4 (Web Text))

Max Activating Token Index: 351

Text #18

Max Range: 0.9531. Min Range: -0.9531

Max Act: 0.7517. Min Act: -0.0001

Data Index: 1170835 (C4 (Web Text))

Max Activating Token Index: 125

Text #19

Max Range: 0.9531. Min Range: -0.9531

Max Act: 0.7337. Min Act: -0.0001

Data Index: 611093 (C4 (Web Text))

Max Activating Token Index: 165