Neuron Explorer

Qwen3-0.6B

capital N starting a name

A neuron in layer 27 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 responds to a capital N that starts a name or word, such as Naperlou, Noreen, Nynke or NRC. It goes negative on the odd lowercase bits that follow a capital letter, like '/N' or '-n'. Turning it up made the model repeat words like 'office' and 'Celsius', and turning it down made the text wander.

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

    need to take time and look around. It is amazing what the world has to teach us!⏎Yes, I am also fully agree with you both (Rob and Naperlou). In this case we can take the best example from nature i

  2. #2 peak

    Anesthetized mice had up to a 25 percent reduction in heart function measures within 20 minutes of exposure to the chemical. Study co-author Nipavan Chiamvimonvat remarked that in this case, triclos

  3. #3 peak

    a Russian scientist, N.V. Lazarev, who was interested in substances that helped the body adapt

  4. #4 peak

    components, and for good reason considering all the essential vitamins and nutrients found in milk. Calcium is one of the most well known ones, but according to Dr. Noreen Iftikhar from the healthline newsletter milk also contains magnesium,

  5. #5 peak

    of billions of dollars, but as the following discoveries prove, the hidden secrets of the seabed are about much more than just money.⏎Billion-dollar wreck: Nuestra Señora de Atocha⏎In July 1985

  6. #6 peak

    human driving the vehicle, but with the rise of automation, the role of the human driver will slowly disappear. Our author takes a closer look.⏎An article by Nynke E Vellinga⏎Two international road traffic Conventions, namely the

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: Oneoneonononononononononononononononononononononononononononononononononononononononononononon

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

    insurance. About 44% of these people were white, 33% were Hispanic, 15% were black, 5% were Asian/Native Hawaiian or Pacific Islander, and 3% were of another race,

  2. #2 peak

    “Being there for those in difficulty, for refugees and migrants, the poor and the lonely and for the victims and perpetrators of violence and abuse, is a non-negotiable requirement if we are to be faithful to

  3. #3 peak

    , which give rise to expectations. Frustrated expectations result in an effective response. Music only has meaning unto itself and the music listener ascribes any extra-musical meaning. This includes any emotional meaning. The unfolding of a song is

  4. #4 peak

    to protect the earth’s climate. The global climate crisis is non-negotiable. #Move the Date: We need to do everything we can to

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. L22:2299 excites
  2. L17:391 excites
  3. L16:55 excites
  4. L5:499 excites
  5. L2:1923 suppresses
  6. L18:127 excites
  7. L19:595 excites
  8. L18:282 excites
  9. L0:210 excites
  10. L10:431 excites
  11. L20:212 excites
  12. L21:2959 excites
  13. L23:1532 suppresses
  14. L3:410 suppresses
  15. L26:1564 excites
  16. L1:27 excites
  17. L18:2453 suppresses
  18. L10:1745 weakly excites
  19. L16:898 weakly suppresses
  20. L0:1484 weakly suppresses
  21. L7:421 weakly suppresses
  22. L2:238 weakly excites
  23. L23:206 weakly suppresses
  24. L11:969 weakly suppresses

Feeds later layers

None: this is the last layer.

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 -30.33 and -27.842 values between -12.91 and -10.431 values between -10.43 and -7.941 values between -7.94 and -5.458 values between -5.45 and -2.9682 values between -2.96 and -0.473789 values between -0.47 and 2.018 values between 2.01 and 4.5011 values between 4.50 and 6.991 values between 9.48 and 11.971 values between 11.97 and 14.451 values between 14.45 and 16.941 values between 19.43 and 21.921 values between 66.70 and 69.19offnegativepositiveoften

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. ense
  2. typingsSlinky
  3. 张
  4. 发放
  5. _shuffle
  6. ChangedEventArgs
  7. EncodingException
  8. ----------------------------------------------------------------------
  9. (records
  10. 纪

Suppresses

  1. e
  2. onn
  3. on
  4. lius
  5. onne
  6. ll
  7. ON
  8. u
  9. onis
  10. eos