Neuron Explorer

Qwen3-0.6B

'are' after a plural subject

A neuron in layer 23 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 word 'are' when it follows a plural noun, as in 'The next largest cities are' or 'Other country sites are'. It is a verb-agreement signal and does not show clearly in the text the model wrote when it was turned up.

Written by Claude from this neuron's activations and steering outputs. Confidence: low. 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

    fact, the largest city in the state, Charleston (the state capital) is home to just 49,736 people. The next largest cities are Huntington (48,638

  2. #2 peak

    at certain times, all the digits are⏎consecutive. How many times like this are there between midnight⏎and 7 a.m.?⏎These practical challenges are all about making a 'tray' and covering it with paper.⏎My cousin

  3. #3 peak

    You can also select a web site from the following list:⏎Select the China site (in Chinese or English) for best site performance. Other MathWorks country sites are not optimized for visits from your location.⏎Contact your local office<|endoftext|>It’s Pr

  4. #4 peak

    You can also select a web site from the following list:⏎Select the China site (in Chinese or English) for best site performance. Other MathWorks country sites are not optimized for visits from your location.⏎Contact your local office<|endoftext|>The point in

  5. #5 peak

    view this ebook on an iPhone, iPad or Android mobile device you will need the Adobe Digital Editions app, or BlueFire Reader or Txtr app. These are free, too. For more details see this article.⏎|Size: ||3

  6. #6 peak

    discouraged by the seemingly complicated written structure. With Music Theory for Dummies, understanding music has never been easier!⏎Note: CD-ROM/DVD and other supplementary materials are not included as part of eBook file.<|endoftext|>is a former municipality in Akers

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: it can is a word that is it is it is it is it is it is it is it is it is it is it is it is it is it is it is it is it is it is it is it is it is it

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

    . . . And Young,” by Lt. Gen. Harold G. Moore and war correspondent Joseph Galloway. The 2002 movie, “We Were Soldiers,” starring Mel Gibson, was based on the Moore-Galloway book

  2. #2 peak

    OTHEISM⏎WE BELIEVE that MONOTHEISM, (TOWHID) has

  3. #3 peak

    about the Python-list<|endoftext|>Jayne Davis: “I am not convinced in my area of research there are less females choosing to take their career in this direction.”⏎We spoke to Jay

  4. #4 peak

    Intelligence. The Director of the Central Intelligence Agency (CIA) is appointed by the President with the advice and consent of the Senate.⏎Who manages the CIA?⏎We are pleased to announce that Bill Burns has been sworn in as the new Director of

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. L22:309 strongly excites
  2. L19:1317 strongly excites
  3. L21:468 strongly excites
  4. L20:1423 strongly suppresses
  5. L19:1422 excites
  6. L18:444 suppresses
  7. L19:1629 excites
  8. L18:1096 excites
  9. L21:2936 suppresses
  10. L8:253 suppresses
  11. L20:1799 excites
  12. L19:545 excites
  13. L2:149 suppresses
  14. L4:450 excites
  15. L16:341 suppresses
  16. L19:880 suppresses
  17. L2:326 suppresses
  18. L4:298 excites
  19. L21:262 excites
  20. L12:642 excites
  21. L17:945 suppresses
  22. L18:1686 excites
  23. L17:147 suppresses
  24. L10:661 suppresses

Feeds later layers

  1. L24:600 strongly suppresses
  2. L26:72 strongly suppresses 'they' or 'we' after where or who
  3. L24:401 strongly excites 'it' referring back to a noun
  4. L24:1080 strongly suppresses
  5. L25:586 strongly excites
  6. L24:241 strongly suppresses
  7. L27:650 strongly suppresses
  8. L24:120 excites
  9. L27:321 excites
  10. L27:500 excites
  11. L26:2947 suppresses
  12. L26:467 excites
  13. L27:461 excites
  14. L25:2055 suppresses
  15. L26:1097 excites
  16. L24:277 excites
  17. L26:1417 suppresses
  18. L26:93 excites
  19. L25:347 excites
  20. L27:577 suppresses
  21. L26:874 suppresses
  22. L27:1256 excites
  23. L25:108 suppresses
  24. L26:926 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 -26.17 and -25.101 values between -22.95 and -21.882 values between -20.80 and -19.731 values between -19.73 and -18.661 values between -18.66 and -17.583 values between -15.43 and -14.367 values between -14.36 and -13.295 values between -13.29 and -12.212 values between -12.21 and -11.145 values between -11.14 and -10.065 values between -10.06 and -8.996 values between -8.99 and -7.9215 values between -7.92 and -6.8411 values between -6.84 and -5.7726 values between -5.77 and -4.7043 values between -4.70 and -3.6288 values between -3.62 and -2.55216 values between -2.55 and -1.47482 values between -1.47 and -0.401831 values between -0.40 and 0.67541 values between 0.67 and 1.75257 values between 1.75 and 2.82130 values between 2.82 and 3.9097 values between 3.90 and 4.9750 values between 4.97 and 6.0427 values between 6.04 and 7.1223 values between 7.12 and 8.198 values between 8.19 and 9.2611 values between 9.26 and 10.347 values between 10.34 and 11.411 values between 11.41 and 12.491 values between 12.49 and 13.562 values between 14.63 and 15.712 values between 15.71 and 16.78offnegativepositiveoften

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. $is
  2. is
  3. meets
  4. grows
  5. survives
  6. does
  7. was
  8. doesn
  9. was
  10. goes

Suppresses

  1. 都可以
  2. 都有
  3. 都能
  4. 都是
  5. 都要
  6. 都
  7. 均可
  8. 都非常
  9. 都很
  10. are