Model: GELU Model: 4 Layers, 2048 Neurons per Layer
Dataset: 80% C4 (Web Text) and 20% Python Code
Neuron 39 in Layer 3
Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-4l')
Text #0
Max Range: 4.2035. Min Range: -4.2035
Max Act: 4.2035. Min Act: -0.1700
Data Index: 5238 (C4 (Web Text))
Max Activating Token Index: 735
Click toggle to see full text
Truncated
Full Text #0
Text #1
Max Range: 4.2035. Min Range: -4.2035
Max Act: 4.1723. Min Act: -0.1700
Data Index: 495561 (C4 (Web Text))
Max Activating Token Index: 1006
Click toggle to see full text
Truncated
Full Text #1
Text #2
Max Range: 4.2035. Min Range: -4.2035
Max Act: 4.2035. Min Act: -0.1700
Data Index: 83631 (C4 (Web Text))
Max Activating Token Index: 536
Click toggle to see full text
Truncated
Full Text #2
Text #3
Max Range: 4.2035. Min Range: -4.2035
Max Act: 4.1098. Min Act: -0.1700
Data Index: 67041 (C4 (Web Text))
Max Activating Token Index: 559
Click toggle to see full text
Truncated
Full Text #3
Text #4
Max Range: 4.2035. Min Range: -4.2035
Max Act: 4.1410. Min Act: -0.1700
Data Index: 883630 (C4 (Web Text))
Max Activating Token Index: 25
Click toggle to see full text
Truncated
Full Text #4
Text #5
Max Range: 4.2035. Min Range: -4.2035
Max Act: 4.0785. Min Act: -0.1700
Data Index: 465014 (C4 (Web Text))
Max Activating Token Index: 132
Click toggle to see full text
Truncated
Full Text #5
Text #6
Max Range: 4.2035. Min Range: -4.2035
Max Act: 4.0785. Min Act: -0.1700
Data Index: 1043958 (C4 (Web Text))
Max Activating Token Index: 636
Click toggle to see full text
Truncated
Full Text #6
Text #7
Max Range: 4.2035. Min Range: -4.2035
Max Act: 4.0472. Min Act: -0.1700
Data Index: 582551 (C4 (Web Text))
Max Activating Token Index: 480
Click toggle to see full text
Truncated
Full Text #7
Text #8
Max Range: 4.2035. Min Range: -4.2035
Max Act: 4.0160. Min Act: -0.1700
Data Index: 880497 (C4 (Web Text))
Max Activating Token Index: 171
Click toggle to see full text
Truncated
Full Text #8
Text #9
Max Range: 4.2035. Min Range: -4.2035
Max Act: 3.9847. Min Act: -0.1700
Data Index: 576431 (C4 (Web Text))
Max Activating Token Index: 734
Click toggle to see full text
Truncated
Full Text #9
Text #10
Max Range: 4.2035. Min Range: -4.2035
Max Act: 3.9847. Min Act: -0.1700
Data Index: 281818 (C4 (Web Text))
Max Activating Token Index: 716
Click toggle to see full text
Truncated
Full Text #10
Text #11
Max Range: 4.2035. Min Range: -4.2035
Max Act: 3.9534. Min Act: -0.1700
Data Index: 1043093 (C4 (Web Text))
Max Activating Token Index: 817
Click toggle to see full text
Truncated
Full Text #11
Text #12
Max Range: 4.2035. Min Range: -4.2035
Max Act: 3.9534. Min Act: -0.1699
Data Index: 457938 (C4 (Web Text))
Max Activating Token Index: 950
Click toggle to see full text
Truncated
Full Text #12
Text #13
Max Range: 4.2035. Min Range: -4.2035
Max Act: 3.9847. Min Act: -0.1700
Data Index: 375453 (C4 (Web Text))
Max Activating Token Index: 853
Click toggle to see full text
Truncated
Full Text #13
Text #14
Max Range: 4.2035. Min Range: -4.2035
Max Act: 3.9065. Min Act: -0.1700
Data Index: 306636 (C4 (Web Text))
Max Activating Token Index: 820
Click toggle to see full text
Truncated
Full Text #14
Text #15
Max Range: 4.2035. Min Range: -4.2035
Max Act: 3.9065. Min Act: -0.1700
Data Index: 1164268 (C4 (Web Text))
Max Activating Token Index: 127
Click toggle to see full text
Truncated
Full Text #15
Text #16
Max Range: 4.2035. Min Range: -4.2035
Max Act: 3.8284. Min Act: -0.1700
Data Index: 448530 (C4 (Web Text))
Max Activating Token Index: 867
Click toggle to see full text
Truncated
Full Text #16
Text #17
Max Range: 4.2035. Min Range: -4.2035
Max Act: 3.8284. Min Act: -0.1700
Data Index: 771422 (C4 (Web Text))
Max Activating Token Index: 551
Click toggle to see full text
Truncated
Full Text #17
Text #18
Max Range: 4.2035. Min Range: -4.2035
Max Act: 3.8753. Min Act: -0.1700
Data Index: 1031500 (C4 (Web Text))
Max Activating Token Index: 767
Click toggle to see full text
Truncated
Full Text #18
Text #19
Max Range: 4.2035. Min Range: -4.2035
Max Act: 3.8284. Min Act: -0.1700
Data Index: 65687 (C4 (Web Text))
Max Activating Token Index: 945
Click toggle to see full text
Truncated
Full Text #19