Model: SoLU Model: 12 Layers, 6144 Neurons per Layer
Dataset: 80% C4 (Web Text) and 20% Python Code
Neuron 4096 in Layer 3
Transformer Lens Loading: HookedTransformer.from_pretrained('solu-12l')
Text #0
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.7951. Min Act: -0.0001
Data Index: 164653 (C4 (Web Text))
Max Activating Token Index: 884
Click toggle to see full text
Truncated
Full Text #0
Text #1
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.8303. Min Act: -0.0001
Data Index: 578900 (C4 (Web Text))
Max Activating Token Index: 987
Click toggle to see full text
Truncated
Full Text #1
Text #2
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.5052. Min Act: -0.0001
Data Index: 1030974 (C4 (Web Text))
Max Activating Token Index: 25
Click toggle to see full text
Truncated
Full Text #2
Text #3
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.5853. Min Act: -0.0001
Data Index: 223540 (C4 (Web Text))
Max Activating Token Index: 382
Click toggle to see full text
Truncated
Full Text #3
Text #4
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.4380. Min Act: -0.0001
Data Index: 316310 (C4 (Web Text))
Max Activating Token Index: 381
Click toggle to see full text
Truncated
Full Text #4
Text #5
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.4521. Min Act: -0.0001
Data Index: 1335637 (C4 (Web Text))
Max Activating Token Index: 614
Click toggle to see full text
Truncated
Full Text #5
Text #6
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.4802. Min Act: -0.0002
Data Index: 431630 (C4 (Web Text))
Max Activating Token Index: 619
Click toggle to see full text
Truncated
Full Text #6
Text #7
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.4209. Min Act: -0.0001
Data Index: 302947 (C4 (Web Text))
Max Activating Token Index: 476
Click toggle to see full text
Truncated
Full Text #7
Text #8
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.3655. Min Act: -0.0001
Data Index: 340871 (C4 (Web Text))
Max Activating Token Index: 412
Click toggle to see full text
Truncated
Full Text #8
Text #9
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.1562. Min Act: -0.0001
Data Index: 447398 (C4 (Web Text))
Max Activating Token Index: 267
Click toggle to see full text
Truncated
Full Text #9
Text #10
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.3022. Min Act: -0.0001
Data Index: 627747 (C4 (Web Text))
Max Activating Token Index: 958
Click toggle to see full text
Truncated
Full Text #10
Text #11
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.2647. Min Act: -0.0002
Data Index: 272684 (C4 (Web Text))
Max Activating Token Index: 698
Click toggle to see full text
Truncated
Full Text #11
Text #12
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.3309. Min Act: -0.0001
Data Index: 825533 (C4 (Web Text))
Max Activating Token Index: 165
Click toggle to see full text
Truncated
Full Text #12
Text #13
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.2358. Min Act: -0.0001
Data Index: 72917 (C4 (Web Text))
Max Activating Token Index: 304
Click toggle to see full text
Truncated
Full Text #13
Text #14
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.2210. Min Act: -0.0001
Data Index: 113694 (C4 (Web Text))
Max Activating Token Index: 220
Click toggle to see full text
Truncated
Full Text #14
Text #15
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.3319. Min Act: -0.0001
Data Index: 1286252 (C4 (Web Text))
Max Activating Token Index: 777
Click toggle to see full text
Truncated
Full Text #15
Text #16
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.0227. Min Act: -0.0001
Data Index: 7647 (C4 (Web Text))
Max Activating Token Index: 716
Click toggle to see full text
Truncated
Full Text #16
Text #17
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.4228. Min Act: -0.0001
Data Index: 163451 (C4 (Web Text))
Max Activating Token Index: 239
Click toggle to see full text
Truncated
Full Text #17
Text #18
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.4188. Min Act: -0.0002
Data Index: 238733 (C4 (Web Text))
Max Activating Token Index: 724
Click toggle to see full text
Truncated
Full Text #18
Text #19
Max Range: 1.8303. Min Range: -1.8303
Max Act: 1.3574. Min Act: -0.0001
Data Index: 506081 (C4 (Web Text))
Max Activating Token Index: 685
Click toggle to see full text
Truncated
Full Text #19