Model: GELU Model: 4 Layers, 2048 Neurons per Layer
Dataset: 80% C4 (Web Text) and 20% Python Code
Neuron 48 in Layer 2
Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-4l')
Text #0
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.7987. Min Act: -0.1700
Data Index: 272982 (C4 (Web Text))
Max Activating Token Index: 461
Click toggle to see full text
Truncated
Full Text #0
Text #1
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.7828. Min Act: -0.1700
Data Index: 361051 (C4 (Web Text))
Max Activating Token Index: 599
Click toggle to see full text
Truncated
Full Text #1
Text #2
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.7987. Min Act: -0.1700
Data Index: 368316 (C4 (Web Text))
Max Activating Token Index: 661
Click toggle to see full text
Truncated
Full Text #2
Text #3
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.7349. Min Act: -0.1700
Data Index: 2187 (C4 (Web Text))
Max Activating Token Index: 875
Click toggle to see full text
Truncated
Full Text #3
Text #4
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.7190. Min Act: -0.1700
Data Index: 1646977 (Python Code)
Max Activating Token Index: 726
Click toggle to see full text
Truncated
Full Text #4
Text #5
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.7349. Min Act: -0.1700
Data Index: 749812 (C4 (Web Text))
Max Activating Token Index: 956
Click toggle to see full text
Truncated
Full Text #5
Text #6
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.6709. Min Act: -0.1700
Data Index: 287086 (C4 (Web Text))
Max Activating Token Index: 976
Click toggle to see full text
Truncated
Full Text #6
Text #7
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.6870. Min Act: -0.1700
Data Index: 1274414 (C4 (Web Text))
Max Activating Token Index: 723
Click toggle to see full text
Truncated
Full Text #7
Text #8
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.7190. Min Act: -0.1700
Data Index: 673885 (C4 (Web Text))
Max Activating Token Index: 825
Click toggle to see full text
Truncated
Full Text #8
Text #9
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.7030. Min Act: -0.1700
Data Index: 91314 (C4 (Web Text))
Max Activating Token Index: 816
Click toggle to see full text
Truncated
Full Text #9
Text #10
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.7030. Min Act: -0.1700
Data Index: 757505 (C4 (Web Text))
Max Activating Token Index: 274
Click toggle to see full text
Truncated
Full Text #10
Text #11
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.7030. Min Act: -0.1700
Data Index: 1126617 (C4 (Web Text))
Max Activating Token Index: 283
Click toggle to see full text
Truncated
Full Text #11
Text #12
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.7190. Min Act: -0.1700
Data Index: 1321091 (C4 (Web Text))
Max Activating Token Index: 1007
Click toggle to see full text
Truncated
Full Text #12
Text #13
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.7030. Min Act: -0.1700
Data Index: 610774 (C4 (Web Text))
Max Activating Token Index: 701
Click toggle to see full text
Truncated
Full Text #13
Text #14
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.7030. Min Act: -0.1700
Data Index: 735280 (C4 (Web Text))
Max Activating Token Index: 388
Click toggle to see full text
Truncated
Full Text #14
Text #15
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.6870. Min Act: -0.1700
Data Index: 924272 (C4 (Web Text))
Max Activating Token Index: 348
Click toggle to see full text
Truncated
Full Text #15
Text #16
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.7030. Min Act: -0.1700
Data Index: 769398 (C4 (Web Text))
Max Activating Token Index: 670
Click toggle to see full text
Truncated
Full Text #16
Text #17
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.6549. Min Act: -0.1700
Data Index: 266427 (C4 (Web Text))
Max Activating Token Index: 770
Click toggle to see full text
Truncated
Full Text #17
Text #18
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.7030. Min Act: -0.1700
Data Index: 956822 (C4 (Web Text))
Max Activating Token Index: 640
Click toggle to see full text
Truncated
Full Text #18
Text #19
Max Range: 2.7987. Min Range: -2.7987
Max Act: 2.7030. Min Act: -0.1700
Data Index: 1643924 (Python Code)
Max Activating Token Index: 585
Click toggle to see full text
Truncated
Full Text #19