Neuron Explorer

Qwen3-0.6B

unclear: 'Le' and 'Gu' name fragments, 'custom' when pushed up

A neuron in layer 26 of Qwen3-0.6B. Its strongest snippets are very consistent: they mostly peak on the same token. It is featured, so it has steering outputs.

What this neuron does

This neuron fires on the middle of a word that is split in two, such as the 'Le' in 'Leblanc' or the 'Gu' in 'Guiana'. When it was turned up, the model kept writing the word 'custom' over and over.

Written by Claude from this neuron's activations and steering outputs. Confidence: medium. Think it is wrong? Flag it (opens a prefilled issue on GitHub).

Fires most on

The text windows where this neuron's activation was highest. Amber marks positive activation; the outlined token is the peak. Hover a token for its strength.

  1. #1 peak

    , the first of Europe's Automated Transfer Vehicles, was launched from the ESA spaceport in French Guiana early March 9. It linked up with the space station almost a month

  2. #2 peak

    , 15th May 1823 as ‘Holbein’⏎Archibald Alexander Leslie Melville, 13th Earl of Leven and 12th Earl of Melville⏎by whom sold, Robinson

  3. #3 peak

    portrait collection following an international public appeal between 1917 and 1920 which was advertised in both the press and ration books.⏎Diane Lees, director general at the IWM, said: "Everybody in the world

  4. #4 peak

    . Against his advice, the pair had married before the Express left Edmonton House. He wondered what the Governor would think.⏎Nancy 'Matooskie' Leblanc had her own reasons for reflection on that hasty marriage. It reminded

  5. #5 peak

    , too, have wildly shifted over the years. Jacques Chirac, at least according to Schofield, was ‘a dyed-in-the-wool Guallist’, and an ideological successor to a young

  6. #6 peak

    between disparate research communities will create new opportunities to advance conservation,” said co-author Dr Allison Leidner of NASA’s Earth Science Division. “For example, it will help

Steering

Push this neuron up or down while the model writes, and watch what changes. The outputs were generated ahead of time, so nothing runs in your browser.

turned down a lot
Prompt

Here is something I have been thinking about lately: how can a first-year student insaps a first-year insaps insaps insaps insaps insaps insaps insaps insaps insaps insaps insaps insaps insaps insaps insaps insaps insaps insaps ins

Most negative on

Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.

  1. #1 peak

    What type of Treatment is best?⏎Every teen with social issues may have a different combination of needs. We meet our clients where they are, assessing their unique situation and tailoring treatment to be the most efficient and effective while encouraging them throughout.<|endoftext|>

  2. #2 peak

    content is as effective as possible.⏎Additionally, AI can help you find your target audience. This can allow you to create specifically designed content to meet your target audience’s needs. Finally, AI can help you

  3. #3 peak

    the barbarous or the savage stage, which displayed a regard for the happiness and well-being of others.⏎It is often difficult to tell how much of Indian manners and morals may have been derived from white people; but there are still some tribes which

  4. #4 peak

    By knowing what works best for me, I can develop learning strategies for myself that play to my own strengths and preferences.⏎Perhaps I can

Connections

Show on map

Which neurons in earlier layers feed this one, and which later ones it feeds. Amber links excite, blue links suppress; longer bars are stronger.

Fed by earlier layers

  1. L22:1396 excites
  2. L23:2284 excites
  3. L23:1933 weakly suppresses
  4. L24:2029 weakly suppresses
  5. L17:2691 weakly suppresses
  6. L25:1400 weakly excites
  7. L15:2160 weakly excites
  8. L11:3020 weakly excites
  9. L21:97 weakly suppresses
  10. L9:445 weakly suppresses
  11. L17:2119 weakly suppresses
  12. L25:943 weakly suppresses
  13. L22:1645 weakly excites
  14. L3:2566 weakly suppresses
  15. L19:2185 weakly excites
  16. L3:2484 weakly suppresses
  17. L21:250 weakly suppresses
  18. L20:2765 weakly suppresses
  19. L12:2471 weakly suppresses
  20. L21:1627 weakly suppresses
  21. L22:678 weakly suppresses
  22. L9:33 weakly suppresses
  23. L5:428 weakly excites
  24. L24:2263 weakly suppresses

Feeds later layers

  1. L27:794 weakly suppresses
  2. L27:2024 weakly suppresses
  3. L27:1863 weakly excites
  4. L27:2339 weakly excites
  5. L27:1753 weakly excites
  6. L27:1868 weakly excites
  7. L27:1310 weakly excites
  8. L27:1498 weakly excites
  9. L27:393 weakly excites
  10. L27:1915 weakly excites
  11. L27:1629 weakly suppresses
  12. L27:2698 weakly suppresses
  13. L27:1763 weakly excites
  14. L27:728 weakly suppresses
  15. L27:814 weakly suppresses
  16. L27:1712 weakly suppresses
  17. L27:2241 weakly suppresses
  18. L27:2798 weakly excites
  19. L27:2107 weakly excites
  20. L27:1983 weakly suppresses
  21. L27:237 weakly suppresses
  22. L27:2815 weakly suppresses
  23. L27:1123 weakly excites
  24. L27:1811 weakly suppresses

Activation histogram

How often this neuron fires at each strength. Almost always it does close to nothing (the tall bars in the middle); the snippets above are the rare moments it fired hard, amber for positive and blue for negative.

1 values between -9.16 and -8.541 values between -8.54 and -7.912 values between -6.66 and -6.032 values between -6.03 and -5.401 values between -5.40 and -4.781 values between -3.52 and -2.901 values between -2.90 and -2.279 values between -2.27 and -1.6432 values between -1.64 and -1.02438 values between -1.02 and -0.393039 values between -0.39 and 0.24342 values between 0.24 and 0.8614 values between 0.86 and 1.495 values between 1.49 and 2.123 values between 2.12 and 2.743 values between 3.37 and 4.002 values between 4.00 and 4.632 values between 4.63 and 5.254 values between 5.25 and 5.881 values between 7.13 and 7.762 values between 8.39 and 9.011 values between 9.01 and 9.641 values between 14.03 and 14.651 values between 15.28 and 15.91offnegativepositiveoften

Logit lens

The words this neuron nudges the model toward (amber) or away from (blue) when it fires. A hint, not a label: many neurons work through later layers instead.

Promotes

  1. Custom
  2. custom
  3. custom
  4. (custom
  5. Custom
  6. customized
  7. .custom
  8. _custom
  9. 定制
  10. .Custom

Suppresses

  1. atorium
  2. tings
  3. .mainloop
  4. �
  5. iture
  6. etchup
  7. mitted
  8. ющ
  9. rix
  10. 饔