Model: SoLU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 173 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-3l')

Text #0

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0369. Min Act: -0.0006

Data Index: 928399 (C4 (Web Text))

Max Activating Token Index: 71

Text #1

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0362. Min Act: -0.0006

Data Index: 197786 (C4 (Web Text))

Max Activating Token Index: 772

Text #2

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0362. Min Act: -0.0006

Data Index: 62542 (C4 (Web Text))

Max Activating Token Index: 33

Text #3

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0357. Min Act: -0.0006

Data Index: 293326 (C4 (Web Text))

Max Activating Token Index: 124

Text #4

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0345. Min Act: -0.0006

Data Index: 252938 (C4 (Web Text))

Max Activating Token Index: 510

Text #5

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0361. Min Act: -0.0006

Data Index: 733611 (C4 (Web Text))

Max Activating Token Index: 812

Text #6

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0345. Min Act: -0.0006

Data Index: 842620 (C4 (Web Text))

Max Activating Token Index: 219

Text #7

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0358. Min Act: -0.0006

Data Index: 1076485 (C4 (Web Text))

Max Activating Token Index: 355

Text #8

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0351. Min Act: -0.0006

Data Index: 197548 (C4 (Web Text))

Max Activating Token Index: 639

Text #9

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0364. Min Act: -0.0006

Data Index: 617208 (C4 (Web Text))

Max Activating Token Index: 147

Text #10

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0323. Min Act: -0.0006

Data Index: 1312703 (C4 (Web Text))

Max Activating Token Index: 724

Text #11

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0346. Min Act: -0.0006

Data Index: 712109 (C4 (Web Text))

Max Activating Token Index: 306

Text #12

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0343. Min Act: -0.0006

Data Index: 1185157 (C4 (Web Text))

Max Activating Token Index: 1012

Text #13

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0349. Min Act: -0.0006

Data Index: 1390181 (Python Code)

Max Activating Token Index: 409

Text #14

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0320. Min Act: -0.0006

Data Index: 974875 (C4 (Web Text))

Max Activating Token Index: 822

Text #15

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0330. Min Act: -0.0006

Data Index: 1085444 (C4 (Web Text))

Max Activating Token Index: 45

Text #16

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0350. Min Act: -0.0006

Data Index: 495958 (C4 (Web Text))

Max Activating Token Index: 201

Text #17

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0327. Min Act: -0.0006

Data Index: 841275 (C4 (Web Text))

Max Activating Token Index: 166

Text #18

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0307. Min Act: -0.0006

Data Index: 140182 (C4 (Web Text))

Max Activating Token Index: 590

Text #19

Max Range: 0.0369. Min Range: -0.0369

Max Act: 0.0304. Min Act: -0.0006

Data Index: 1292543 (C4 (Web Text))

Max Activating Token Index: 650