Model: GELU Model: 3 Layers, 2048 Neurons per Layer
Dataset: 80% C4 (Web Text) and 20% Python Code
Neuron 2019 in Layer 1
Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-3l')
Text #0
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.6601. Min Act: -0.1700
Data Index: 632266 (C4 (Web Text))
Max Activating Token Index: 403
Click toggle to see full text
Truncated
Full Text #0
Text #1
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.7539. Min Act: -0.1700
Data Index: 677857 (C4 (Web Text))
Max Activating Token Index: 485
Click toggle to see full text
Truncated
Full Text #1
Text #2
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.5976. Min Act: -0.1700
Data Index: 1256946 (C4 (Web Text))
Max Activating Token Index: 150
Click toggle to see full text
Truncated
Full Text #2
Text #3
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.5976. Min Act: -0.1700
Data Index: 281310 (C4 (Web Text))
Max Activating Token Index: 532
Click toggle to see full text
Truncated
Full Text #3
Text #4
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.5663. Min Act: -0.1700
Data Index: 795096 (C4 (Web Text))
Max Activating Token Index: 374
Click toggle to see full text
Truncated
Full Text #4
Text #5
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.5351. Min Act: -0.1700
Data Index: 3917 (C4 (Web Text))
Max Activating Token Index: 309
Click toggle to see full text
Truncated
Full Text #5
Text #6
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.5038. Min Act: -0.1700
Data Index: 1085405 (C4 (Web Text))
Max Activating Token Index: 300
Click toggle to see full text
Truncated
Full Text #6
Text #7
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.5038. Min Act: -0.1700
Data Index: 138600 (C4 (Web Text))
Max Activating Token Index: 647
Click toggle to see full text
Truncated
Full Text #7
Text #8
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.4726. Min Act: -0.1700
Data Index: 481667 (C4 (Web Text))
Max Activating Token Index: 184
Click toggle to see full text
Truncated
Full Text #8
Text #9
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.5351. Min Act: -0.1700
Data Index: 1258087 (C4 (Web Text))
Max Activating Token Index: 660
Click toggle to see full text
Truncated
Full Text #9
Text #10
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.4413. Min Act: -0.1700
Data Index: 918568 (C4 (Web Text))
Max Activating Token Index: 125
Click toggle to see full text
Truncated
Full Text #10
Text #11
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.4413. Min Act: -0.1700
Data Index: 973118 (C4 (Web Text))
Max Activating Token Index: 720
Click toggle to see full text
Truncated
Full Text #11
Text #12
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.4413. Min Act: -0.1700
Data Index: 328065 (C4 (Web Text))
Max Activating Token Index: 271
Click toggle to see full text
Truncated
Full Text #12
Text #13
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.4726. Min Act: -0.1700
Data Index: 46005 (C4 (Web Text))
Max Activating Token Index: 957
Click toggle to see full text
Truncated
Full Text #13
Text #14
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.4413. Min Act: -0.1700
Data Index: 101228 (C4 (Web Text))
Max Activating Token Index: 506
Click toggle to see full text
Truncated
Full Text #14
Text #15
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.4101. Min Act: -0.1700
Data Index: 853543 (C4 (Web Text))
Max Activating Token Index: 257
Click toggle to see full text
Truncated
Full Text #15
Text #16
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.3788. Min Act: -0.1700
Data Index: 196885 (C4 (Web Text))
Max Activating Token Index: 555
Click toggle to see full text
Truncated
Full Text #16
Text #17
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.3788. Min Act: -0.1700
Data Index: 473945 (C4 (Web Text))
Max Activating Token Index: 122
Click toggle to see full text
Truncated
Full Text #17
Text #18
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.3788. Min Act: -0.1700
Data Index: 1194099 (C4 (Web Text))
Max Activating Token Index: 360
Click toggle to see full text
Truncated
Full Text #18
Text #19
Max Range: 4.7539. Min Range: -4.7539
Max Act: 4.4726. Min Act: -0.1700
Data Index: 26495 (C4 (Web Text))
Max Activating Token Index: 172
Click toggle to see full text
Truncated
Full Text #19