Neuron Explorer

Qwen3-0.6B

'Sch' in surnames

A neuron in layer 26 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 fires on the letters 'Sch' when they start a name, such as Schinske, Schenk or Schaeffer. Turned up, the model began writing strings of curly quote marks and the symbol '《', which fits what the neuron tends to push toward.

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

    no one ethnic or racial group forming a majority of students. During a plenary presentation at the conference, Jeff Schinske, professor of biology at De Anza Community College, and three current

  2. #2 peak

    2- to 4-Year Institutions⏎Shifting Students’ Stereotypes of Scientists to Enhance Science Identity in a Diverse, Community College Context⏎Jeff Schinske, Amanda Snyder, Heather Perkins, Mary Wyner, Monica Cardenas

  3. #3 peak

    radio service in Model X. Because AM audio quality can be very poor, we offer internet radio to give<|endoftext|>A New Brunswick Ghost Story⏎S. E. Schlosser⏎It was a cold winter and the snow was deep in the woods

  4. #4 peak

    Notes⏎Lilium 'White Stargazer' is the most popular of the white upright lilies and used by many for forcing. In fact, Peter Schenk bred it with forcing and the cut flower trade in mind. However, L

  5. #5 peak

    K, Cortese M, Healy BC, Kuhle J, Mina MJ, Leng Y, Elledge SJ, Niebuhr DW, Scher AI, Munger KL and Ascherio A⏎Multiple sclerosis (MS

  6. #6 peak

    of teaching older kids the value of money is by giving them an allowance. You can make a weekly schedules of chores that they are responsible for around<|endoftext|>by Cole Schenewerk - The California

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: what would be a good approach to make really good really good really really really really really really really really really really really really really really really really really really really really really really really really really really really really really really really really really really really really

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

    “You have to read this,” you do it. I’m usually not a fan of historical fiction, particularly novels set during wars, but Joseph Bruchac’s Code Talker is an exception, a book I’m so happy to have been

  2. #2 peak

    TV shows such as “Prometheus” and “Star Trek” are all the more fascinating because of

  3. #3 peak

    - blue paper⏎- black paper⏎- pencil top eraser⏎The piece we selected to recreate was Van Gogh’s Starry Night. We’ve been learning about apples this week, an easy tie

  4. #4 peak

    government because it is lower than what we believe to be the actual price inflation rate and makes GDP look better than it really is.<|endoftext|>For the genre including games like Final Fantasy and Summon Night:

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. L25:386 strongly suppresses
  2. L24:1601 strongly suppresses
  3. L24:405 strongly excites
  4. L23:1333 strongly suppresses
  5. L21:2300 strongly excites
  6. L24:989 excites
  7. L20:1406 excites
  8. L24:256 suppresses
  9. L25:1440 excites
  10. L20:2918 excites
  11. L18:2293 suppresses
  12. L20:2412 suppresses
  13. L21:2926 excites
  14. L21:881 suppresses comma after 'if' clause in prose, braces in code
  15. L23:292 suppresses
  16. L21:2216 suppresses
  17. L19:1979 excites
  18. L22:2362 excites
  19. L22:460 excites blank spot right before a number
  20. L22:1973 excites
  21. L24:880 excites
  22. L23:1692 excites
  23. L22:470 excites
  24. L23:486 excites

Feeds later layers

  1. L27:967 strongly excites
  2. L27:1324 excites
  3. L27:1045 suppresses
  4. L27:1165 excites
  5. L27:315 excites
  6. L27:697 excites
  7. L27:2964 excites
  8. L27:1255 weakly excites
  9. L27:2506 weakly excites
  10. L27:772 weakly excites
  11. L27:1759 weakly excites
  12. L27:661 weakly excites
  13. L27:1779 weakly excites
  14. L27:267 weakly excites
  15. L27:770 weakly excites
  16. L27:252 weakly excites
  17. L27:1942 weakly excites
  18. L27:444 weakly excites
  19. L27:554 weakly excites
  20. L27:234 weakly excites
  21. L27:615 weakly excites
  22. L27:806 weakly excites
  23. L27:1120 weakly excites
  24. L27:1099 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 -5.34 and -5.041 values between -3.53 and -3.231 values between -2.93 and -2.631 values between -2.63 and -2.336 values between -2.33 and -2.037 values between -2.03 and -1.7320 values between -1.73 and -1.4348 values between -1.43 and -1.13191 values between -1.13 and -0.82668 values between -0.82 and -0.521710 values between -0.52 and -0.221130 values between -0.22 and 0.0892 values between 0.08 and 0.3812 values between 0.38 and 0.683 values between 0.68 and 0.984 values between 0.98 and 1.282 values between 1.28 and 1.593 values between 1.59 and 1.891 values between 1.89 and 2.192 values between 2.19 and 2.492 values between 3.39 and 3.691 values between 3.99 and 4.302 values between 6.40 and 6.70offnegativepositiveoften

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. 《
  2. ${
  3. ,《
  4. `{
  5. ({
  6. ${
  7. {'
  8. ({
  9. {"
  10. {

Suppresses

  1. uries
  2. iasi
  3. apia
  4. ncy
  5. otron
  6. ฆ
  7. ArrayType
  8. vandalism
  9. пров
  10. stinian