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

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

Neuron 3423 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.3051. Min Act: 0.0000

Data Index: 650205 (C4 (Web Text))

Max Activating Token Index: 496

Text #1

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.2689. Min Act: -0.0000

Data Index: 91519 (C4 (Web Text))

Max Activating Token Index: 572

Text #2

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.2802. Min Act: -0.0000

Data Index: 757990 (C4 (Web Text))

Max Activating Token Index: 568

Text #3

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.2651. Min Act: -0.0000

Data Index: 244668 (C4 (Web Text))

Max Activating Token Index: 201

Text #4

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.2722. Min Act: -0.0000

Data Index: 1466164 (Python Code)

Max Activating Token Index: 683

Text #5

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.2687. Min Act: -0.0000

Data Index: 850011 (C4 (Web Text))

Max Activating Token Index: 113

Text #6

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.2564. Min Act: -0.0000

Data Index: 473063 (C4 (Web Text))

Max Activating Token Index: 220

Text #7

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.2780. Min Act: -0.0000

Data Index: 279922 (C4 (Web Text))

Max Activating Token Index: 282

Text #8

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.2659. Min Act: -0.0000

Data Index: 978810 (C4 (Web Text))

Max Activating Token Index: 165

Text #9

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.2656. Min Act: -0.0000

Data Index: 876435 (C4 (Web Text))

Max Activating Token Index: 730

Text #10

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.2621. Min Act: -0.0000

Data Index: 483003 (C4 (Web Text))

Max Activating Token Index: 216

Text #11

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.2534. Min Act: -0.0000

Data Index: 1587177 (Python Code)

Max Activating Token Index: 546

Text #12

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.2605. Min Act: -0.0000

Data Index: 1547865 (Python Code)

Max Activating Token Index: 737

Text #13

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.2593. Min Act: -0.0000

Data Index: 929743 (C4 (Web Text))

Max Activating Token Index: 323

Text #14

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.2511. Min Act: 0.0000

Data Index: 1082161 (C4 (Web Text))

Max Activating Token Index: 303

Text #15

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.2487. Min Act: -0.0000

Data Index: 75790 (C4 (Web Text))

Max Activating Token Index: 693

Text #16

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.2741. Min Act: -0.0000

Data Index: 710592 (C4 (Web Text))

Max Activating Token Index: 833

Text #17

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.2620. Min Act: -0.0000

Data Index: 1138077 (C4 (Web Text))

Max Activating Token Index: 545

Text #18

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.2460. Min Act: -0.0000

Data Index: 469337 (C4 (Web Text))

Max Activating Token Index: 102

Text #19

Max Range: 0.3051. Min Range: -0.3051

Max Act: 0.2599. Min Act: -0.0000

Data Index: 409389 (C4 (Web Text))

Max Activating Token Index: 146