Neuron Explorer

Qwen3-0.6B

first piece of split proper names

A neuron in layer 26 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 the first chunk of a name or place that the model has split into pieces, such as the "Or" in Origen or the "Merr" in Merrimack. It goes quiet, and slightly negative, on words like "it" or "but" that come after a comma or a clause. Turning it up made the text drift into odd repeats of the word "need" or "light", which does not clearly connect to names.

Written by Claude from this neuron's activations and steering outputs. Confidence: medium. Think it is wrong? Flag it (opens an email to the maker with the neuron already filled in).

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

    ) testifies (Adv. Har. xxix.1) It is derived from the Hebrew Ebion, "poor," and was not given, as Origen supposes, for their low view of Christ. Id. at 6

  2. #2 peak

    . However, right around 235, Hippolytus was reconciled with the Church.⏎One of the other items on Pontian’s plate was Origen. As a young man, Origen established a school of philosophy in Alexandria

  3. #3 peak

    Saint-Galmier (Loire).⏎Seven years later, he found a partnership with Bessac and the painter Joséphus Veyrat. Then, he set up his own business in Grigny,

  4. #4 peak

    Boulard can offer gingivitis treatment in Merrimack, NH. When left undiagnosed, gingivitis can

  5. #5 peak

    Once you develop a familiarity with topics such as the difference between regression and classification models, you’ll be able to solve an array of machine learning problems. Author Scott V. Burger provides several examples to help you build a working knowledge of machine learning.⏎

  6. #6 peak

    Phormidium pseudpriestleyi.⏎Phormidium pseudpriestleyi isolated from a meltwater pond on the McMurdo Ice Shelf, Antarctica© Patricia Mesquita⏎Microbial mat with Ph

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: light light light light light light light light light light light light light light light light light light light light light light light light light light light light light light light light light light light light light light light light light light light light light light light light

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

    is based on objective investigation, reasoning and analysis as opposed to subjectivity and prejudice, belief and myth. Evidence may be derived from a purely scientific investigation, but it has to be allied to the political, cultural, economic and social dimensions of these

  2. #2 peak

    . Think fresh fruits and vegetables, nuts, seeds, oils, and whole grains. While a whole foods diet does focus on eating lots of plant-based foods, it does not necessarily mean you have to become a vegetarian or vegan. Meat and dairy

  3. #3 peak

    fertilizers and pesticides, extensive soil erosion, pollution, and high rates of greenhouse gas emissions. Low-intensity farming methods on the other hand–which include, but aren’t limited to, organic agriculture–typically favour more environmentally-friendly farming practices at

  4. #4 peak

    quialism, exclamations, repetitions, and circumlocutions—in short, all the features of a mind instantly verbalizing. In this way, the reader experiences with freshness and immediacy the thoughtless ease with which they wound each other

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. L16:466 suppresses 'is' after a long subject phrase
  2. L15:184 suppresses
  3. L25:586 weakly excites
  4. L19:2324 weakly excites
  5. L1:81 weakly excites
  6. L21:262 weakly suppresses
  7. L18:541 weakly suppresses
  8. L21:2710 weakly excites
  9. L21:1657 weakly excites
  10. L23:176 weakly excites
  11. L17:3018 weakly suppresses
  12. L21:456 weakly suppresses
  13. L24:241 weakly excites
  14. L18:233 weakly excites
  15. L19:545 weakly suppresses
  16. L4:450 weakly suppresses
  17. L9:74 weakly suppresses
  18. L22:309 weakly suppresses
  19. L23:287 weakly excites 'are' after a plural subject
  20. L11:840 weakly excites
  21. L24:401 weakly excites 'it' referring back to a noun
  22. L16:184 weakly excites
  23. L19:2398 weakly excites
  24. L19:2002 weakly suppresses 'is' or 'was' after a subject

Feeds later layers

  1. L27:1493 strongly suppresses
  2. L27:1123 suppresses
  3. L27:1202 weakly excites
  4. L27:2598 weakly excites
  5. L27:883 weakly suppresses
  6. L27:1797 weakly suppresses
  7. L27:2576 weakly excites capital N starting a name
  8. L27:1301 weakly excites
  9. L27:1234 weakly excites
  10. L27:935 weakly suppresses
  11. L27:1177 weakly suppresses
  12. L27:2527 weakly excites
  13. L27:2767 weakly suppresses
  14. L27:1030 weakly excites
  15. L27:2845 weakly excites
  16. L27:2635 weakly suppresses
  17. L27:1239 weakly excites
  18. L27:29 weakly excites
  19. L27:310 weakly excites
  20. L27:2698 weakly suppresses
  21. L27:3059 weakly suppresses
  22. L27:758 weakly excites
  23. L27:1771 weakly excites
  24. L27:1514 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 -31.06 and -29.391 values between -27.73 and -26.061 values between -26.06 and -24.392 values between -21.05 and -19.383 values between -19.38 and -17.715 values between -17.71 and -16.045 values between -16.04 and -14.389 values between -14.38 and -12.714 values between -12.71 and -11.046 values between -11.04 and -9.379 values between -9.37 and -7.7016 values between -7.70 and -6.0321 values between -6.03 and -4.3644 values between -4.36 and -2.69108 values between -2.69 and -1.033407 values between -1.03 and 0.64168 values between 0.64 and 2.3142 values between 2.31 and 3.9819 values between 3.98 and 5.658 values between 5.65 and 7.3210 values between 7.32 and 8.996 values between 8.99 and 10.661 values between 12.33 and 13.991 values between 13.99 and 15.664 values between 15.66 and 17.331 values between 17.33 and 19.001 values between 19.00 and 20.672 values between 20.67 and 22.342 values between 27.34 and 29.011 values between 34.02 and 35.69offnegativepositiveoften

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. pio
  2. ARDS
  3. เฉพาะ
  4. rades
  5. からは
  6. odia
  7. rire
  8. ients
  9. marshall
  10. デン

Suppresses

  1. need
  2. need
  3. Need
  4. Need
  5. _need
  6. NEED
  7. 需要
  8. needing
  9. 需要用
  10. want