Neuron Explorer

Qwen3-0.6B

'young' before people or adults

A neuron in layer 24 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 lights up on the word "young" when it comes right before "people" or "adults", as in "young people" or "youth and young adults". When it was turned up, the model's writing drifted into unrelated topics, so the steering does not show a clear effect.

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

    them.⏎But they clearly contain nicotine, which is addictive. Public health advocates warn that e-cigarette makers are using the same strategies that tobacco companies employed to attract young people — candy flavorings, for example. Lax state-level regulations also make

  2. #2 peak

    unfortunately, in the absence of evidence-based public health messages regarding the health risks of e-cigarettes, television advertising is promoting beliefs and behaviors that pose harm to youth and young adults and raise public health concerns.”⏎Researchers analyzed TV advertising data and viewership ratings

  3. #3 peak

    simply believe what they see and read. A recent study at Stanford University into how well people evaluate stories on social media came up with this conclusion; “Overall, young people’s ability to reason about the information on the Internet can be summed up in

  4. #4 peak

    But smaller players are being forced to consolidate to cut costs or close their doors. Consolidation and closures have led to rising unemployment at a time when 2 million young people enter the national workforce each year.⏎“Business is only getting better for the

  5. #5 peak

    a single brand, blu eCigs, which is owned by the tobacco company Lorillard. The study is the first to extensively analyze trends in youth and young adult exposure to e-cigarette TV ads.⏎“If the current trends continue, awareness

  6. #6 peak

    40 percent of college-educated workers make less than the top 25 percent of high school graduates.⏎This alternative reading of the figure helps to explain why many young people have not responded so enthusiastically to the “market signals” about college. On

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: The most important 500-550 550-600 550-650 550-700 550-750 550-

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

    can build apps for different platforms like iOS, Android, or cross-platform frameworks. Proper coding practices ensure that apps are efficient, secure, and scalable, delivering a smooth user experience. With the ever-growing demand for innovative apps, proficient coding skills have

  2. #2 peak

    ).⏎Extension modules play an active role in all these aspects of the Apache server's⏎Figure 3-1. The HTTP transaction consists⏎of a URI request from the browser to the server, followed by a document response⏎from the server to

  3. #3 peak

    and to transform it into plain-vanilla HTML, XHTML, or other formats (PDF, plain text, rich text files, and so forth) for delivery to Web clients. This helps sidestep the kinds of horrible compatibility issues that dynamic

  4. #4 peak

    Or, a fluffly feline sits next to a glass of water, representing "tengo sed," or "I am thirsty."⏎The app also challenges its students to answer increasingly difficult typing, match-the-phrase-to-the-photo, listening

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. L22:2642 strongly excites
  2. L19:1876 excites
  3. L22:1721 excites
  4. L23:1406 suppresses
  5. L17:2528 suppresses
  6. L20:2161 suppresses
  7. L23:98 excites
  8. L20:1436 suppresses
  9. L21:1864 excites
  10. L17:3062 suppresses
  11. L17:3003 excites
  12. L23:436 excites
  13. L19:2840 excites
  14. L17:1518 suppresses
  15. L22:712 suppresses
  16. L16:2041 excites
  17. L19:2845 suppresses
  18. L22:1636 suppresses
  19. L23:1493 excites
  20. L22:94 suppresses
  21. L19:2996 suppresses
  22. L22:496 suppresses
  23. L19:1970 excites
  24. L17:1847 excites

Feeds later layers

  1. L26:657 strongly excites the word 'people' coming next
  2. L27:2339 excites
  3. L25:1560 excites
  4. L27:1515 weakly excites
  5. L27:1884 weakly excites
  6. L25:1291 weakly excites
  7. L27:3069 weakly excites
  8. L26:552 weakly excites
  9. L27:1339 weakly suppresses
  10. L25:707 weakly excites
  11. L25:2707 weakly suppresses
  12. L25:1547 weakly excites
  13. L26:1201 weakly suppresses
  14. L27:1257 weakly excites
  15. L25:173 weakly suppresses
  16. L25:1550 weakly suppresses
  17. L26:663 weakly suppresses
  18. L25:1177 weakly suppresses
  19. L25:2952 weakly excites
  20. L27:2516 weakly excites
  21. L26:2543 weakly excites
  22. L27:1411 weakly excites
  23. L27:3012 weakly suppresses
  24. L26:1127 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.

1 values between -20.02 and -18.782 values between -7.65 and -6.422 values between -6.42 and -5.185 values between -5.18 and -3.9410 values between -3.94 and -2.7123 values between -2.71 and -1.47988 values between -1.47 and -0.232294 values between -0.23 and 1.00277 values between 1.00 and 2.24120 values between 2.24 and 3.4868 values between 3.48 and 4.7136 values between 4.71 and 5.9530 values between 5.95 and 7.1818 values between 7.18 and 8.4212 values between 8.42 and 9.664 values between 9.66 and 10.892 values between 10.89 and 12.135 values between 12.13 and 13.373 values between 13.37 and 14.602 values between 14.60 and 15.842 values between 15.84 and 17.072 values between 17.07 and 18.312 values between 28.20 and 29.44offnegativepositiveoften

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. people
  2. person
  3. people
  4. People
  5. PEOPLE
  6. persons
  7. person
  8. People
  9. _people
  10. (person

Suppresses

  1. eken
  2. 候选
  3. göre
  4. player
  5. roker
  6. Player
  7. .float
  8. ecute
  9. .components
  10. ajaran