Model: SoLU Model: 8 Layers, 4096 Neurons per Layer
Dataset: 80% C4 (Web Text) and 20% Python Code
Neuron 2016 in Layer 3
Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')
Text #0
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.4991. Min Act: -0.0001
Data Index: 686928 (C4 (Web Text))
Max Activating Token Index: 127
Click toggle to see full text
Truncated
Full Text #0
Text #1
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.4604. Min Act: -0.0001
Data Index: 1148319 (C4 (Web Text))
Max Activating Token Index: 600
Click toggle to see full text
Truncated
Full Text #1
Text #2
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.4242. Min Act: -0.0001
Data Index: 61945 (C4 (Web Text))
Max Activating Token Index: 917
Click toggle to see full text
Truncated
Full Text #2
Text #3
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.3722. Min Act: -0.0001
Data Index: 1050534 (C4 (Web Text))
Max Activating Token Index: 748
Click toggle to see full text
Truncated
Full Text #3
Text #4
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.3012. Min Act: -0.0001
Data Index: 1266568 (C4 (Web Text))
Max Activating Token Index: 264
Click toggle to see full text
Truncated
Full Text #4
Text #5
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.2664. Min Act: -0.0001
Data Index: 1387059 (Python Code)
Max Activating Token Index: 221
Click toggle to see full text
Truncated
Full Text #5
Text #6
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.2958. Min Act: -0.0001
Data Index: 349375 (C4 (Web Text))
Max Activating Token Index: 299
Click toggle to see full text
Truncated
Full Text #6
Text #7
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.2838. Min Act: -0.0001
Data Index: 254213 (C4 (Web Text))
Max Activating Token Index: 450
Click toggle to see full text
Truncated
Full Text #7
Text #8
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.2308. Min Act: -0.0001
Data Index: 1400972 (Python Code)
Max Activating Token Index: 184
Click toggle to see full text
Truncated
Full Text #8
Text #9
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.2647. Min Act: -0.0001
Data Index: 1365555 (Python Code)
Max Activating Token Index: 508
Click toggle to see full text
Truncated
Full Text #9
Text #10
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.2243. Min Act: -0.0001
Data Index: 1572114 (Python Code)
Max Activating Token Index: 111
Click toggle to see full text
Truncated
Full Text #10
Text #11
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.2069. Min Act: -0.0001
Data Index: 848884 (C4 (Web Text))
Max Activating Token Index: 889
Click toggle to see full text
Truncated
Full Text #11
Text #12
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.1495. Min Act: -0.0001
Data Index: 1175325 (C4 (Web Text))
Max Activating Token Index: 923
Click toggle to see full text
Truncated
Full Text #12
Text #13
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.2255. Min Act: -0.0001
Data Index: 1447369 (Python Code)
Max Activating Token Index: 196
Click toggle to see full text
Truncated
Full Text #13
Text #14
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.2139. Min Act: -0.0001
Data Index: 474534 (C4 (Web Text))
Max Activating Token Index: 681
Click toggle to see full text
Truncated
Full Text #14
Text #15
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.2489. Min Act: -0.0001
Data Index: 1188865 (C4 (Web Text))
Max Activating Token Index: 939
Click toggle to see full text
Truncated
Full Text #15
Text #16
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.1805. Min Act: -0.0001
Data Index: 1033331 (C4 (Web Text))
Max Activating Token Index: 918
Click toggle to see full text
Truncated
Full Text #16
Text #17
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.2361. Min Act: -0.0001
Data Index: 664501 (C4 (Web Text))
Max Activating Token Index: 85
Click toggle to see full text
Truncated
Full Text #17
Text #18
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.1932. Min Act: -0.0001
Data Index: 1499329 (Python Code)
Max Activating Token Index: 231
Click toggle to see full text
Truncated
Full Text #18
Text #19
Max Range: 1.4991. Min Range: -1.4991
Max Act: 1.1585. Min Act: -0.0001
Data Index: 1546615 (Python Code)
Max Activating Token Index: 82
Click toggle to see full text
Truncated
Full Text #19