Model: GELU Model: 3 Layers, 2048 Neurons per Layer
Dataset: 80% C4 (Web Text) and 20% Python Code
Neuron 813 in Layer 0
Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-3l')
Text #0
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.9591. Min Act: -0.1700
Data Index: 7178 (C4 (Web Text))
Max Activating Token Index: 424
Click toggle to see full text
Truncated
Full Text #0
Text #1
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.9275. Min Act: -0.1700
Data Index: 1641663 (Python Code)
Max Activating Token Index: 599
Click toggle to see full text
Truncated
Full Text #1
Text #2
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.9116. Min Act: -0.1700
Data Index: 1004810 (C4 (Web Text))
Max Activating Token Index: 354
Click toggle to see full text
Truncated
Full Text #2
Text #3
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.9116. Min Act: -0.1700
Data Index: 815242 (C4 (Web Text))
Max Activating Token Index: 693
Click toggle to see full text
Truncated
Full Text #3
Text #4
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.8957. Min Act: -0.1700
Data Index: 308078 (C4 (Web Text))
Max Activating Token Index: 676
Click toggle to see full text
Truncated
Full Text #4
Text #5
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.8640. Min Act: -0.1700
Data Index: 798834 (C4 (Web Text))
Max Activating Token Index: 530
Click toggle to see full text
Truncated
Full Text #5
Text #6
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.8957. Min Act: -0.1700
Data Index: 307486 (C4 (Web Text))
Max Activating Token Index: 983
Click toggle to see full text
Truncated
Full Text #6
Text #7
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.8799. Min Act: -0.1700
Data Index: 965119 (C4 (Web Text))
Max Activating Token Index: 700
Click toggle to see full text
Truncated
Full Text #7
Text #8
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.8640. Min Act: -0.1700
Data Index: 1643312 (Python Code)
Max Activating Token Index: 3
Click toggle to see full text
Truncated
Full Text #8
Text #9
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.8481. Min Act: -0.1700
Data Index: 1553717 (Python Code)
Max Activating Token Index: 387
Click toggle to see full text
Truncated
Full Text #9
Text #10
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.8799. Min Act: -0.1700
Data Index: 937182 (C4 (Web Text))
Max Activating Token Index: 293
Click toggle to see full text
Truncated
Full Text #10
Text #11
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.8640. Min Act: -0.1700
Data Index: 472765 (C4 (Web Text))
Max Activating Token Index: 25
Click toggle to see full text
Truncated
Full Text #11
Text #12
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.8322. Min Act: -0.1700
Data Index: 83805 (C4 (Web Text))
Max Activating Token Index: 616
Click toggle to see full text
Truncated
Full Text #12
Text #13
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.8481. Min Act: -0.1700
Data Index: 807618 (C4 (Web Text))
Max Activating Token Index: 189
Click toggle to see full text
Truncated
Full Text #13
Text #14
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.8322. Min Act: -0.1700
Data Index: 1201241 (C4 (Web Text))
Max Activating Token Index: 634
Click toggle to see full text
Truncated
Full Text #14
Text #15
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.8322. Min Act: -0.1700
Data Index: 686949 (C4 (Web Text))
Max Activating Token Index: 906
Click toggle to see full text
Truncated
Full Text #15
Text #16
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.8322. Min Act: -0.1700
Data Index: 805658 (C4 (Web Text))
Max Activating Token Index: 958
Click toggle to see full text
Truncated
Full Text #16
Text #17
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.8322. Min Act: -0.1700
Data Index: 440239 (C4 (Web Text))
Max Activating Token Index: 483
Click toggle to see full text
Truncated
Full Text #17
Text #18
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.8163. Min Act: -0.1700
Data Index: 924957 (C4 (Web Text))
Max Activating Token Index: 873
Click toggle to see full text
Truncated
Full Text #18
Text #19
Max Range: 2.9591. Min Range: -2.9591
Max Act: 2.8004. Min Act: -0.1700
Data Index: 520127 (C4 (Web Text))
Max Activating Token Index: 701
Click toggle to see full text
Truncated
Full Text #19