Model: GELU Model: 4 Layers, 2048 Neurons per Layer
Dataset: 80% C4 (Web Text) and 20% Python Code
Neuron 1994 in Layer 1
Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-4l')
Text #0
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.8416. Min Act: -0.1700
Data Index: 1166276 (C4 (Web Text))
Max Activating Token Index: 771
Click toggle to see full text
Truncated
Full Text #0
Text #1
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.7779. Min Act: -0.1700
Data Index: 165901 (C4 (Web Text))
Max Activating Token Index: 21
Click toggle to see full text
Truncated
Full Text #1
Text #2
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.7939. Min Act: -0.1700
Data Index: 1195508 (C4 (Web Text))
Max Activating Token Index: 269
Click toggle to see full text
Truncated
Full Text #2
Text #3
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.7301. Min Act: -0.1700
Data Index: 1124603 (C4 (Web Text))
Max Activating Token Index: 966
Click toggle to see full text
Truncated
Full Text #3
Text #4
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.7141. Min Act: -0.1700
Data Index: 653895 (C4 (Web Text))
Max Activating Token Index: 178
Click toggle to see full text
Truncated
Full Text #4
Text #5
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.7620. Min Act: -0.1700
Data Index: 118689 (C4 (Web Text))
Max Activating Token Index: 846
Click toggle to see full text
Truncated
Full Text #5
Text #6
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.6981. Min Act: -0.1700
Data Index: 1356990 (C4 (Web Text))
Max Activating Token Index: 991
Click toggle to see full text
Truncated
Full Text #6
Text #7
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.6981. Min Act: -0.1700
Data Index: 31824 (C4 (Web Text))
Max Activating Token Index: 26
Click toggle to see full text
Truncated
Full Text #7
Text #8
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.7141. Min Act: -0.1700
Data Index: 381560 (C4 (Web Text))
Max Activating Token Index: 244
Click toggle to see full text
Truncated
Full Text #8
Text #9
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.7141. Min Act: -0.1700
Data Index: 477736 (C4 (Web Text))
Max Activating Token Index: 473
Click toggle to see full text
Truncated
Full Text #9
Text #10
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.6981. Min Act: -0.1700
Data Index: 102956 (C4 (Web Text))
Max Activating Token Index: 477
Click toggle to see full text
Truncated
Full Text #10
Text #11
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.6981. Min Act: -0.1700
Data Index: 1546260 (Python Code)
Max Activating Token Index: 321
Click toggle to see full text
Truncated
Full Text #11
Text #12
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.6339. Min Act: -0.1700
Data Index: 33922 (C4 (Web Text))
Max Activating Token Index: 998
Click toggle to see full text
Truncated
Full Text #12
Text #13
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.6821. Min Act: -0.1700
Data Index: 1200054 (C4 (Web Text))
Max Activating Token Index: 886
Click toggle to see full text
Truncated
Full Text #13
Text #14
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.6660. Min Act: -0.1700
Data Index: 68104 (C4 (Web Text))
Max Activating Token Index: 324
Click toggle to see full text
Truncated
Full Text #14
Text #15
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.6821. Min Act: -0.1700
Data Index: 942688 (C4 (Web Text))
Max Activating Token Index: 820
Click toggle to see full text
Truncated
Full Text #15
Text #16
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.6660. Min Act: -0.1700
Data Index: 65879 (C4 (Web Text))
Max Activating Token Index: 369
Click toggle to see full text
Truncated
Full Text #16
Text #17
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.6821. Min Act: -0.1700
Data Index: 941504 (C4 (Web Text))
Max Activating Token Index: 935
Click toggle to see full text
Truncated
Full Text #17
Text #18
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.6500. Min Act: -0.1700
Data Index: 702279 (C4 (Web Text))
Max Activating Token Index: 504
Click toggle to see full text
Truncated
Full Text #18
Text #19
Max Range: 2.8416. Min Range: -2.8416
Max Act: 2.6500. Min Act: -0.1700
Data Index: 769150 (C4 (Web Text))
Max Activating Token Index: 260
Click toggle to see full text
Truncated
Full Text #19