Neuron Explorer

Qwen3-0.6B

'going' or 'go' after a noun or 'is'

A neuron in layer 20 of Qwen3-0.6B. Its strongest snippets are fairly consistent: several of them peak on the same token. It is featured, so it has steering outputs.

What this neuron does

This neuron fires on words such as 'going', 'go' and 'true' that follow a subject or a form of 'is', as in 'the ones who are going to live' or 'courage to go'. The snippets are loosely related, and turning the neuron up gave text that did not clearly connect to the prompts.

Written by Claude from this neuron's activations and steering outputs. Confidence: low. 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

    another smash hit would break through in American entertainment. Film director Spike Lee’s “Do The Right Thing” was set in the same multicultural Brooklyn as “The Cosby Show,” but

  2. #2 peak

    failure. Think of it as time-released success.⏎21. There are no failures. Just experiences and your reactions to them.⏎22. Courage is going from failure to failure without losing enthusiasm.⏎23. Failure is unimportant.

  3. #3 peak

    most of the range and the crests of the adjacent San Gabriel range have lost their kiss of snow, big San Gorgonio often still shines white—the one true jewel

  4. #4 peak

    are paying attention to what is happening at the COP16 climate negotiations in Cancun, Mexico. They are the “climate generation” — the ones who are going to have to live with our climate legacy, our tendency to prevaricate and

  5. #5 peak

    fearful aspects of being small in a big world. Bravery is having the courage to go out every day and tackle what comes, no matter what comes. These creatively structured

  6. #6 peak

    Discoveries class on code.org, jump into CircuitPython to learn Python and hardware together, TinyGO, or even use the Arduino IDE. Circuit Playground Express is the newest and best Circuit

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: It is possible to show that the same between the same is to show that between the same is to show that between the same is to show that between the same is to show that between the same is to show that between the same is to

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

    and air conditioning systems, it also means that pollutants that originate indoors are getting trapped inside.⏎“People normally<|endoftext|>Spring Dawn by Meng Haoran 春眠不觉晓 处处闻啼茑 夜来风雨声 花落

  2. #2 peak

    221BC). Unlike most Chinese idioms, 毋忘在莒 strangely has actually two stories of origin, which gave this idiom two

  3. #3 peak

    Nihon Shoki 日本書紀 (720) 9 and in the Hizen-no-kuni fudoki 肥前國風土記 (first part of the 8th century) 10. The

  4. #4 peak

    become the Duke of Qi and started building up the state with reform and dedication. Years passed on since his days as a helpless young heir in the state of Ju, he started to really enjoy the good old

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. L18:125 suppresses
  2. L15:59 suppresses
  3. L11:223 weakly excites
  4. L19:28 weakly suppresses
  5. L2:49 weakly suppresses
  6. L18:66 weakly suppresses
  7. L10:155 weakly excites
  8. L4:2662 weakly suppresses
  9. L16:1308 weakly excites
  10. L8:328 weakly excites
  11. L12:2243 weakly suppresses
  12. L16:513 weakly excites
  13. L4:607 weakly suppresses
  14. L13:44 weakly suppresses inside web addresses and citations
  15. L4:16 weakly excites
  16. L12:1628 weakly excites
  17. L1:2922 weakly excites
  18. L5:2115 weakly excites
  19. L7:2217 weakly suppresses
  20. L13:1848 weakly excites
  21. L16:2225 weakly suppresses
  22. L9:1129 weakly excites
  23. L18:1156 weakly excites
  24. L0:1697 weakly suppresses

Feeds later layers

  1. L24:1920 weakly suppresses
  2. L26:2449 weakly excites
  3. L22:2073 weakly excites
  4. L22:1075 weakly suppresses
  5. L23:582 weakly suppresses
  6. L24:320 weakly excites
  7. L22:2900 weakly suppresses
  8. L21:2041 weakly excites
  9. L24:1650 weakly suppresses
  10. L22:2619 weakly excites
  11. L24:2584 weakly excites 'in' in set phrases like 'in mind'
  12. L21:1652 weakly suppresses
  13. L25:2255 weakly excites
  14. L21:2856 weakly excites
  15. L21:1267 weakly suppresses
  16. L25:1725 weakly excites
  17. L25:1354 weakly suppresses
  18. L21:1034 weakly excites
  19. L23:1994 weakly excites
  20. L25:2056 weakly suppresses
  21. L25:2991 weakly excites
  22. L21:2670 weakly excites
  23. L23:1626 weakly suppresses
  24. L26:787 weakly excites

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.

7 values between -1.24 and -1.0220 values between -1.02 and -0.8184 values between -0.81 and -0.59270 values between -0.59 and -0.38693 values between -0.38 and -0.161448 values between -0.16 and 0.05898 values between 0.05 and 0.27365 values between 0.27 and 0.4883 values between 0.48 and 0.7020 values between 0.70 and 0.916 values between 0.91 and 1.131 values between 1.13 and 1.351 values between 1.35 and 1.563 values between 1.56 and 1.783 values between 1.78 and 1.991 values between 1.99 and 2.211 values between 2.21 and 2.421 values between 2.42 and 2.642 values between 2.64 and 2.851 values between 7.16 and 7.37offnegativepositiveoften

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. plete
  2. ative
  3. dictions
  4. mnop
  5. tega
  6. urrence
  7. utive
  8. like
  9. iko
  10. missive

Suppresses

  1. 之
  2. 之間
  3. 之间的
  4. 之人
  5. 之间
  6. 起来
  7. 起來
  8. arasında
  9. 之乡
  10. 之内