Model: GELU Model: 4 Layers, 2048 Neurons per Layer
Dataset: 80% C4 (Web Text) and 20% Python Code
Neuron 640 in Layer 2
Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-4l')
Text #0
Max Range: 3.0499. Min Range: -3.0499
Max Act: 3.0499. Min Act: -0.1700
Data Index: 81680 (C4 (Web Text))
Max Activating Token Index: 20
Click toggle to see full text
Truncated
Full Text #0
Text #1
Max Range: 3.0499. Min Range: -3.0499
Max Act: 2.9392. Min Act: -0.1700
Data Index: 1019276 (C4 (Web Text))
Max Activating Token Index: 348
Click toggle to see full text
Truncated
Full Text #1
Text #2
Max Range: 3.0499. Min Range: -3.0499
Max Act: 2.9551. Min Act: -0.1700
Data Index: 974937 (C4 (Web Text))
Max Activating Token Index: 364
Click toggle to see full text
Truncated
Full Text #2
Text #3
Max Range: 3.0499. Min Range: -3.0499
Max Act: 2.9551. Min Act: -0.1700
Data Index: 1145855 (C4 (Web Text))
Max Activating Token Index: 70
Click toggle to see full text
Truncated
Full Text #3
Text #4
Max Range: 3.0499. Min Range: -3.0499
Max Act: 2.8758. Min Act: -0.1700
Data Index: 1540517 (Python Code)
Max Activating Token Index: 270
Click toggle to see full text
Truncated
Full Text #4
Text #5
Max Range: 3.0499. Min Range: -3.0499
Max Act: 2.9234. Min Act: -0.1700
Data Index: 240785 (C4 (Web Text))
Max Activating Token Index: 427
Click toggle to see full text
Truncated
Full Text #5
Text #6
Max Range: 3.0499. Min Range: -3.0499
Max Act: 2.8281. Min Act: -0.1700
Data Index: 438891 (C4 (Web Text))
Max Activating Token Index: 422
Click toggle to see full text
Truncated
Full Text #6
Text #7
Max Range: 3.0499. Min Range: -3.0499
Max Act: 2.8440. Min Act: -0.1700
Data Index: 1463566 (Python Code)
Max Activating Token Index: 893
Click toggle to see full text
Truncated
Full Text #7
Text #8
Max Range: 3.0499. Min Range: -3.0499
Max Act: 2.8758. Min Act: -0.1700
Data Index: 1487006 (Python Code)
Max Activating Token Index: 233
Click toggle to see full text
Truncated
Full Text #8
Text #9
Max Range: 3.0499. Min Range: -3.0499
Max Act: 2.8599. Min Act: -0.1700
Data Index: 1420514 (Python Code)
Max Activating Token Index: 813
Click toggle to see full text
Truncated
Full Text #9
Text #10
Max Range: 3.0499. Min Range: -3.0499
Max Act: 2.8281. Min Act: -0.1700
Data Index: 269799 (C4 (Web Text))
Max Activating Token Index: 453
Click toggle to see full text
Truncated
Full Text #10
Text #11
Max Range: 3.0499. Min Range: -3.0499
Max Act: 2.8440. Min Act: -0.1700
Data Index: 1416346 (Python Code)
Max Activating Token Index: 511
Click toggle to see full text
Truncated
Full Text #11
Text #12
Max Range: 3.0499. Min Range: -3.0499
Max Act: 2.8917. Min Act: -0.1700
Data Index: 1504682 (Python Code)
Max Activating Token Index: 387
Click toggle to see full text
Truncated
Full Text #12
Text #13
Max Range: 3.0499. Min Range: -3.0499
Max Act: 2.8440. Min Act: -0.1700
Data Index: 1360658 (Python Code)
Max Activating Token Index: 280
Click toggle to see full text
Truncated
Full Text #13
Text #14
Max Range: 3.0499. Min Range: -3.0499
Max Act: 2.8281. Min Act: -0.1700
Data Index: 835560 (C4 (Web Text))
Max Activating Token Index: 12
Click toggle to see full text
Truncated
Full Text #14
Text #15
Max Range: 3.0499. Min Range: -3.0499
Max Act: 2.8281. Min Act: -0.1700
Data Index: 924602 (C4 (Web Text))
Max Activating Token Index: 282
Click toggle to see full text
Truncated
Full Text #15
Text #16
Max Range: 3.0499. Min Range: -3.0499
Max Act: 2.7963. Min Act: -0.1700
Data Index: 1515725 (Python Code)
Max Activating Token Index: 188
Click toggle to see full text
Truncated
Full Text #16
Text #17
Max Range: 3.0499. Min Range: -3.0499
Max Act: 2.8122. Min Act: -0.1700
Data Index: 1596003 (Python Code)
Max Activating Token Index: 213
Click toggle to see full text
Truncated
Full Text #17
Text #18
Max Range: 3.0499. Min Range: -3.0499
Max Act: 2.7804. Min Act: -0.1700
Data Index: 465744 (C4 (Web Text))
Max Activating Token Index: 886
Click toggle to see full text
Truncated
Full Text #18
Text #19
Max Range: 3.0499. Min Range: -3.0499
Max Act: 2.7963. Min Act: -0.1700
Data Index: 563510 (C4 (Web Text))
Max Activating Token Index: 713
Click toggle to see full text
Truncated
Full Text #19