Neuron Explorer

Qwen3-0.6B

"the" right after "without"

A neuron in layer 2 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 "the" when it comes straight after "without", as in "without the protection" or "without the owner's knowledge". When it was turned up, the model's text did not visibly change in a way that connects to this idea.

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

    , and spectacular towers offering strategic vantage points were just some of the mechanisms used back in the day.⏎Now try imagining your business as a castle, just without the adequate protection. It’s undeniable that external threats are targeting businesses of all sizes,

  2. #2 peak

    impact."⏎As a result, he says, "Most houses in North Carolina that are oceanfront will not make it through a 30-year mortgage" without the government spending money to improve and sustain the area's infrastructure. "The sea level

  3. #3 peak

    common purpose. Conversely, traditional work teams are highly structured and are sometimes confused with the work group. Teamwork can be difficult to establish within a traditional team without the identification of a common purpose to which all workers can agree and commit.⏎The self

  4. #4 peak

    dead at my feet than guilty of a mortal sin.” His biographers have written of the long hours he spent in prayer, fasting, and penance, without the knowledge of his people.⏎The French king was an avid lover of justice, who

  5. #5 peak

    appreciate the other side of e-cigarettes. For those unwilling or unable to give up nicotine, e-cigarettes offer something akin to the experience of smoking a real cigarette without the same concentration

  6. #6 peak

    a minimum, they should be cognizant that it is malicious software that collects personal information, usually with the intent of sending that information to third parties, without the owner's knowledge or consent. If anyone would like to learn the historical, technical

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: I have been thinking about lately, and I have been thinking about... But this is not just a new month, but this is not just a new year, but this is not just a new week. I have been thinking about lately,

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

    Link for teaching and learning in the New Year. The ideas are for those who are just starting out, or for those who

  2. #2 peak

    -nine cases cured and a large number "permanently arrested and greatly improved." Von Ruck continued<|endoftext|>Tamales are traditional Mexican foods often served at New Year’s Eve or Christmas Eve dinners. They are steamed bundles of corn dough stuffed

  3. #3 peak

    vegetables and less processed foods combined with Americans’ favorite past-time of adopting restrictive, hard-to-maintain, weight-loss inducing diets at the turn of the New Year, creates a perfect storm for a surge in interest and popularity of the raw food

  4. #4 peak

    of the Jewish high holidays and the civil year on the Jewish calendar. It’s a good time to wish your Jewish friends leshana tova (happy New Year), to send greeting cards, and perhaps plant trees in Israel in their honor.⏎

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. L1:1796 weakly excites
  2. L0:178 weakly suppresses
  3. L0:1301 weakly suppresses
  4. L1:59 weakly excites 'the' after a verb or preposition
  5. L0:1224 weakly suppresses
  6. L0:253 weakly suppresses
  7. L1:24 weakly excites
  8. L1:2689 weakly suppresses
  9. L1:305 weakly suppresses
  10. L1:2188 weakly suppresses
  11. L1:2611 weakly suppresses
  12. L1:2074 weakly suppresses
  13. L1:2831 weakly excites
  14. L1:2119 weakly excites
  15. L0:774 weakly suppresses
  16. L0:510 weakly suppresses
  17. L1:2518 weakly excites
  18. L0:2340 weakly suppresses
  19. L0:2686 weakly suppresses
  20. L0:2999 weakly excites
  21. L1:2919 weakly suppresses
  22. L0:2706 weakly excites
  23. L1:2843 weakly excites
  24. L1:2041 weakly suppresses

Feeds later layers

  1. L16:2621 weakly excites
  2. L18:2336 weakly suppresses
  3. L23:1034 weakly suppresses "set" before out, off or up
  4. L8:2461 weakly excites
  5. L20:2639 weakly suppresses
  6. L17:446 weakly excites
  7. L18:95 weakly suppresses end of a document marker
  8. L15:2600 weakly excites
  9. L17:1909 weakly excites
  10. L9:1532 weakly suppresses
  11. L5:2537 weakly excites
  12. L11:1012 weakly excites
  13. L13:511 weakly suppresses
  14. L11:289 weakly excites
  15. L11:421 weakly suppresses
  16. L6:1209 weakly excites
  17. L14:2573 weakly excites
  18. L4:2171 weakly suppresses
  19. L15:2493 weakly excites
  20. L3:1663 weakly suppresses
  21. L7:2817 weakly excites
  22. L15:2181 weakly suppresses
  23. L11:1471 weakly excites
  24. L8:478 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 -2.94 and -2.821 values between -2.70 and -2.581 values between -2.58 and -2.462 values between -2.34 and -2.224 values between -2.22 and -2.102 values between -2.10 and -1.986 values between -1.98 and -1.864 values between -1.86 and -1.742 values between -1.74 and -1.625 values between -1.62 and -1.505 values between -1.50 and -1.382 values between -1.38 and -1.266 values between -1.26 and -1.143 values between -1.14 and -1.023 values between -0.90 and -0.786 values between -0.78 and -0.664 values between -0.66 and -0.549 values between -0.54 and -0.4227 values between -0.42 and -0.3098 values between -0.30 and -0.18786 values between -0.18 and -0.061892 values between -0.06 and 0.07803 values between 0.07 and 0.19175 values between 0.19 and 0.3134 values between 0.31 and 0.4310 values between 0.43 and 0.557 values between 0.55 and 0.673 values between 0.67 and 0.791 values between 0.79 and 0.911 values between 0.91 and 1.031 values between 1.03 and 1.151 values between 1.27 and 1.391 values between 1.39 and 1.511 values between 1.63 and 1.751 values between 1.75 and 1.87offnegativepositiveoften

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. @foreach
  2. ��
  3. 写字
  4. 资本市场
  5. OWNER
  6. astr
  7. acres
  8. 率达
  9. urai
  10. ':'

Suppresses

  1. énom
  2. wares
  3. rome
  4. ripsi
  5. genu
  6. visor
  7. körper
  8. cludes
  9. ideshow
  10. roids