Model: GELU Model: 4 Layers, 2048 Neurons per Layer
Dataset: 80% C4 (Web Text) and 20% Python Code
Neuron 2005 in Layer 0
Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-4l')
Text #0
Max Range: 3.0188. Min Range: -3.0188
Max Act: 3.0188. Min Act: -0.1700
Data Index: 1313005 (C4 (Web Text))
Max Activating Token Index: 730
Click toggle to see full text
Truncated
Full Text #0
Text #1
Max Range: 3.0188. Min Range: -3.0188
Max Act: 3.0030. Min Act: -0.1700
Data Index: 368120 (C4 (Web Text))
Max Activating Token Index: 272
Click toggle to see full text
Truncated
Full Text #1
Text #2
Max Range: 3.0188. Min Range: -3.0188
Max Act: 2.9397. Min Act: -0.1700
Data Index: 135231 (C4 (Web Text))
Max Activating Token Index: 872
Click toggle to see full text
Truncated
Full Text #2
Text #3
Max Range: 3.0188. Min Range: -3.0188
Max Act: 2.9080. Min Act: -0.1700
Data Index: 1645715 (Python Code)
Max Activating Token Index: 293
Click toggle to see full text
Truncated
Full Text #3
Text #4
Max Range: 3.0188. Min Range: -3.0188
Max Act: 2.9238. Min Act: -0.1700
Data Index: 587065 (C4 (Web Text))
Max Activating Token Index: 111
Click toggle to see full text
Truncated
Full Text #4
Text #5
Max Range: 3.0188. Min Range: -3.0188
Max Act: 2.9080. Min Act: -0.1700
Data Index: 766530 (C4 (Web Text))
Max Activating Token Index: 803
Click toggle to see full text
Truncated
Full Text #5
Text #6
Max Range: 3.0188. Min Range: -3.0188
Max Act: 2.9080. Min Act: -0.1700
Data Index: 205298 (C4 (Web Text))
Max Activating Token Index: 275
Click toggle to see full text
Truncated
Full Text #6
Text #7
Max Range: 3.0188. Min Range: -3.0188
Max Act: 2.8921. Min Act: -0.1700
Data Index: 446113 (C4 (Web Text))
Max Activating Token Index: 762
Click toggle to see full text
Truncated
Full Text #7
Text #8
Max Range: 3.0188. Min Range: -3.0188
Max Act: 2.8286. Min Act: -0.1700
Data Index: 1262571 (C4 (Web Text))
Max Activating Token Index: 116
Click toggle to see full text
Truncated
Full Text #8
Text #9
Max Range: 3.0188. Min Range: -3.0188
Max Act: 2.7808. Min Act: -0.1700
Data Index: 1637408 (Python Code)
Max Activating Token Index: 627
Click toggle to see full text
Truncated
Full Text #9
Text #10
Max Range: 3.0188. Min Range: -3.0188
Max Act: 2.7808. Min Act: -0.1700
Data Index: 1526869 (Python Code)
Max Activating Token Index: 732
Click toggle to see full text
Truncated
Full Text #10
Text #11
Max Range: 3.0188. Min Range: -3.0188
Max Act: 2.7648. Min Act: -0.1700
Data Index: 1548308 (Python Code)
Max Activating Token Index: 344
Click toggle to see full text
Truncated
Full Text #11
Text #12
Max Range: 3.0188. Min Range: -3.0188
Max Act: 2.7489. Min Act: -0.1700
Data Index: 3378 (C4 (Web Text))
Max Activating Token Index: 951
Click toggle to see full text
Truncated
Full Text #12
Text #13
Max Range: 3.0188. Min Range: -3.0188
Max Act: 2.7169. Min Act: -0.1700
Data Index: 1612480 (Python Code)
Max Activating Token Index: 226
Click toggle to see full text
Truncated
Full Text #13
Text #14
Max Range: 3.0188. Min Range: -3.0188
Max Act: 2.7169. Min Act: -0.1700
Data Index: 1630335 (Python Code)
Max Activating Token Index: 866
Click toggle to see full text
Truncated
Full Text #14
Text #15
Max Range: 3.0188. Min Range: -3.0188
Max Act: 2.7329. Min Act: -0.1700
Data Index: 509421 (C4 (Web Text))
Max Activating Token Index: 953
Click toggle to see full text
Truncated
Full Text #15
Text #16
Max Range: 3.0188. Min Range: -3.0188
Max Act: 2.7169. Min Act: -0.1700
Data Index: 424530 (C4 (Web Text))
Max Activating Token Index: 615
Click toggle to see full text
Truncated
Full Text #16
Text #17
Max Range: 3.0188. Min Range: -3.0188
Max Act: 2.7009. Min Act: -0.1700
Data Index: 1054996 (C4 (Web Text))
Max Activating Token Index: 205
Click toggle to see full text
Truncated
Full Text #17
Text #18
Max Range: 3.0188. Min Range: -3.0188
Max Act: 2.7169. Min Act: -0.1700
Data Index: 1384991 (Python Code)
Max Activating Token Index: 163
Click toggle to see full text
Truncated
Full Text #18
Text #19
Max Range: 3.0188. Min Range: -3.0188
Max Act: 2.7169. Min Act: -0.1700
Data Index: 1617517 (Python Code)
Max Activating Token Index: 822
Click toggle to see full text
Truncated
Full Text #19