Neuron Explorer

Qwen3-0.6B

unclear: word closing a two-word term

A neuron in layer 2 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

It fires on the last word of a short phrase or name, like 'circuit' in 'short-circuit', 'Pacific' in 'Great Pacific' or 'track' in 'fast track'. The pattern is loose and the steering outputs don't show a clear theme.

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

    you remember when you learned multiplication and discovered multiplication was a faster and shorter way to represent repeated addition? Well there is a faster and shorter way to represent repeated multiplication.⏎Fractions - Adding Mixed Numbers⏎Learn how to Add

  2. #2 peak

    C. Marshall, who was the US Secretary of State. Thousands of millions of dollars were provided for rebuilding cities, roads, industries etc⏎Definition from the Longman Dictionary of Contemporary English Advanced Learner's Dictionary.⏎Dictionary pictures of the day⏎

  3. #3 peak

    of fault lines circling the Pacific Basin that is prone to frequent earthquakes and volcanic eruptions.⏎CNN's Kathy Quiano contributed to this report.<|endoftext|>From Longman Dictionary of Contemporary Englishldoce_734_znoticeno‧t

  4. #4 peak

    Maximum short-circuit current⏎18. Current acquisition channel⏎19. The internal resistance of

  5. #5 peak

    or short-circuiting. A fuse works by breaking an electrical circuit if the current flowing through the

  6. #6 peak

    to ensure accuracy in their projects. Similarly, many new software development projects keep engineering applications in mind during their development. Young engineers looking to put themselves on the fast track to success would benefit greatly from learning computer technology and science applications relevant to their field

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: " The following is the problem: " A man's in the first place in the second place in the third place in the fourth place in the fifth place, and the sixth place, and the seventh place, and the eighth place, and

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

    or by contacts with immunogens.⏎||Conferred by introduction of ready-made antibodies.⏎||Immune response is durable and effective.⏎||Immune response is short lived and less effective.⏎||Immunity develops only after a lag period.⏎||Imm

  2. #2 peak

    to staying up-to-date, boosting your immune system is wellness 101. Better yet, it’s simpler than you might think (and tastes great, too!). Additionally

  3. #3 peak

    , he became infected three years ago from using dirty needles. Thanks to a daily cocktail of 12 anti-retroviral drugs which stops him developing full-blown AIDS, Vladimir can now live an almost normal life. Today, he spends his

  4. #4 peak

    can disrupt the pathway by binding to certain factors and preventing them from cleaving and activating other factors. In this way, coagulation is circumvented.⏎Sure enough, Simukunin binds to clotting factors and interferes with coag

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. L0:1075 excites
  2. L1:861 suppresses
  3. L0:191 weakly suppresses
  4. L0:2641 weakly suppresses
  5. L1:491 weakly excites
  6. L1:2677 weakly suppresses
  7. L1:1029 weakly excites
  8. L1:647 weakly excites
  9. L0:1301 weakly suppresses
  10. L0:903 weakly excites
  11. L0:178 weakly suppresses
  12. L1:1874 weakly excites
  13. L1:89 weakly excites
  14. L0:2415 weakly suppresses
  15. L0:280 weakly suppresses
  16. L0:1224 weakly suppresses
  17. L1:1340 weakly suppresses
  18. L1:257 weakly suppresses
  19. L1:2291 weakly suppresses
  20. L0:1814 weakly suppresses
  21. L1:2164 weakly suppresses
  22. L1:791 weakly suppresses
  23. L0:2842 weakly suppresses
  24. L1:401 weakly suppresses

Feeds later layers

  1. L4:656 suppresses
  2. L11:922 weakly suppresses
  3. L5:2885 weakly suppresses
  4. L24:1059 weakly excites
  5. L3:844 weakly suppresses
  6. L14:411 weakly excites
  7. L4:723 weakly suppresses
  8. L9:1906 weakly suppresses
  9. L9:1834 weakly suppresses
  10. L13:200 weakly excites
  11. L8:1012 weakly suppresses
  12. L13:2289 weakly excites
  13. L8:825 weakly suppresses
  14. L16:1148 weakly excites
  15. L13:2898 weakly excites
  16. L23:1152 weakly excites
  17. L4:3058 weakly suppresses
  18. L4:1441 weakly suppresses
  19. L11:2994 weakly suppresses
  20. L8:525 weakly suppresses
  21. L7:171 weakly suppresses
  22. L27:979 weakly suppresses
  23. L12:849 weakly excites
  24. L13:1819 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 -2.80 and -2.712 values between -2.61 and -2.511 values between -2.42 and -2.321 values between -1.83 and -1.741 values between -1.64 and -1.541 values between -1.25 and -1.151 values between -1.15 and -1.061 values between -0.96 and -0.861 values between -0.86 and -0.772 values between -0.67 and -0.573 values between -0.57 and -0.482 values between -0.48 and -0.3813 values between -0.38 and -0.2893 values between -0.28 and -0.18408 values between -0.18 and -0.091787 values between -0.09 and 0.011147 values between 0.01 and 0.11319 values between 0.11 and 0.2096 values between 0.20 and 0.3019 values between 0.30 and 0.401 values between 0.40 and 0.493 values between 0.49 and 0.591 values between 0.59 and 0.691 values between 0.69 and 0.791 values between 0.79 and 0.882 values between 0.98 and 1.08offnegativepositiveoften

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. 贯
  2. contenu
  3. jer
  4. казалось
  5. 呗
  6. 钢铁
  7. сет
  8. 煤炭
  9. 权威
  10. 극

Suppresses

  1. ritch
  2. rowser
  3. ancy
  4. ساء
  5. .gc
  6. ortal
  7. وغير
  8. .iOS
  9. ��
  10. doInBackground