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

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

Neuron 4115 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3612. Min Act: -0.0001

Data Index: 706680 (C4 (Web Text))

Max Activating Token Index: 472

Text #1

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3617. Min Act: -0.0001

Data Index: 226910 (C4 (Web Text))

Max Activating Token Index: 319

Text #2

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3735. Min Act: -0.0001

Data Index: 1061797 (C4 (Web Text))

Max Activating Token Index: 976

Text #3

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3723. Min Act: -0.0001

Data Index: 277857 (C4 (Web Text))

Max Activating Token Index: 720

Text #4

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3440. Min Act: -0.0001

Data Index: 1323705 (C4 (Web Text))

Max Activating Token Index: 48

Text #5

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3803. Min Act: -0.0001

Data Index: 551543 (C4 (Web Text))

Max Activating Token Index: 650

Text #6

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3279. Min Act: -0.0001

Data Index: 1251879 (C4 (Web Text))

Max Activating Token Index: 349

Text #7

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3716. Min Act: -0.0001

Data Index: 750275 (C4 (Web Text))

Max Activating Token Index: 756

Text #8

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3335. Min Act: -0.0000

Data Index: 1430129 (Python Code)

Max Activating Token Index: 114

Text #9

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3301. Min Act: -0.0001

Data Index: 122400 (C4 (Web Text))

Max Activating Token Index: 1018

Text #10

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3471. Min Act: -0.0001

Data Index: 1198092 (C4 (Web Text))

Max Activating Token Index: 406

Text #11

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3659. Min Act: -0.0001

Data Index: 1297475 (C4 (Web Text))

Max Activating Token Index: 402

Text #12

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3362. Min Act: -0.0000

Data Index: 117123 (C4 (Web Text))

Max Activating Token Index: 338

Text #13

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3386. Min Act: -0.0000

Data Index: 1389471 (Python Code)

Max Activating Token Index: 784

Text #14

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3388. Min Act: -0.0001

Data Index: 1016802 (C4 (Web Text))

Max Activating Token Index: 703

Text #15

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3394. Min Act: -0.0000

Data Index: 1497368 (Python Code)

Max Activating Token Index: 403

Text #16

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3359. Min Act: -0.0001

Data Index: 71017 (C4 (Web Text))

Max Activating Token Index: 548

Text #17

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3352. Min Act: -0.0001

Data Index: 151115 (C4 (Web Text))

Max Activating Token Index: 369

Text #18

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3301. Min Act: -0.0001

Data Index: 762198 (C4 (Web Text))

Max Activating Token Index: 800

Text #19

Max Range: 0.3803. Min Range: -0.3803

Max Act: 0.3330. Min Act: -0.0001

Data Index: 397168 (C4 (Web Text))

Max Activating Token Index: 510