Neuron Explorer

Qwen3-0.6B

the '0' in years like 2050

A neuron in layer 22 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 zero in future years written as 2 then 0, such as 2050 or 2030, while the model is still reading the number. On the other side it fires negatively on 'in' just before 'the future', and turning it down made the model repeat the word 'future'.

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

    similar study did find that cats decl<|endoftext|>With the human population expected to grow by billions over the new few decades, the United Nations is predicting that by 2050 our food production must double

  2. #2 peak

    , it released a report titled ‘Sharing the effort under a global carbon budget’.⏎WWF says: “A strict global carbon budget between now and 2050 based on a fair distribution between rich and poor nations has

  3. #3 peak

    Climate Change, to forecast sea level changes over the next 30 years. At Duck, the hardest-hit tidal gauge, the panel projects that by 20

  4. #4 peak

    below 2 degrees Celsius.”⏎The report is based on research and shows different ways to cut global emissions by at least 80% globally, by 2050, and by 30% by 2030,

  5. #5 peak

    would have been excessive anyway.⏎According to China’s 5-year plan, their target for coal-powered capacity is 1,100 GW by 2020. But by completing all their power plant projects, they would have hit

  6. #6 peak

    , which includes measures aimed at achieving the president’s goal of net zero emissions by 2050 and a completely carbon-free domestic power industry by 2035.⏎The package may include new tax credits for investments in renewable energy generation

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 future in the future in the future. The future in the future in the future the future in the future in the future in the future the future in the future the future in the future the future in the future in the future in the

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

    an ash cloud over Alaska and lost all power before pilots were able to restart the engines. Ottawa wanted to reassure passengers that steps were being taken to avoid such crises in the future. So did other countries.⏎Twenty years later, the system was working

  2. #2 peak

    the cause of miscarriage remains unknown.⏎Though one in five pregnancies end in miscarriage, having a miscarriage does not make a woman more likely to miscarry in the future.⏎Some people use the word "miscarriage" to describe a

  3. #3 peak

    measures what matters. It tells us how well nations are doing in terms of supporting their inhabitants to live good lives now, while ensuring that others can do the same in the future, i.e

  4. #4 peak

    to research from RateSetter, with 53% doing something which may actually harm their credit score – potentially making it difficult or impossible for them to borrow money in the future.⏎More than half of people

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. L21:2675 strongly excites
  2. L19:553 excites
  3. L19:564 excites
  4. L16:2965 suppresses
  5. L18:3047 suppresses
  6. L21:1699 suppresses
  7. L18:2817 excites
  8. L15:2466 suppresses
  9. L20:2152 suppresses
  10. L18:2532 excites
  11. L14:2915 excites
  12. L14:2691 suppresses
  13. L15:2338 excites
  14. L20:2574 excites
  15. L21:794 suppresses
  16. L14:2516 excites
  17. L19:1999 excites
  18. L18:2178 excites
  19. L19:245 excites
  20. L7:521 suppresses
  21. L21:2426 suppresses
  22. L17:1623 excites
  23. L20:1878 excites
  24. L12:1532 excites

Feeds later layers

  1. L26:2717 strongly suppresses word just before 'future'
  2. L23:1971 strongly suppresses
  3. L25:2440 suppresses
  4. L25:2579 suppresses
  5. L25:1275 weakly suppresses
  6. L24:2353 weakly suppresses
  7. L25:1990 weakly suppresses
  8. L23:240 weakly suppresses
  9. L25:928 weakly suppresses
  10. L24:377 weakly suppresses
  11. L27:2277 weakly suppresses
  12. L26:2782 weakly suppresses
  13. L26:207 weakly excites
  14. L24:1665 weakly excites
  15. L23:1462 weakly excites
  16. L23:201 weakly excites
  17. L26:1578 weakly suppresses
  18. L26:2147 weakly suppresses
  19. L23:457 weakly suppresses
  20. L25:2512 weakly excites
  21. L26:2903 weakly suppresses
  22. L25:1968 weakly suppresses
  23. L23:84 weakly excites
  24. L23:2493 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 -12.68 and -12.121 values between -12.12 and -11.551 values between -10.99 and -10.421 values between -10.42 and -9.862 values between -9.30 and -8.731 values between -8.73 and -8.172 values between -8.17 and -7.612 values between -7.04 and -6.481 values between -6.48 and -5.911 values between -5.91 and -5.352 values between -5.35 and -4.793 values between -4.79 and -4.226 values between -4.22 and -3.669 values between -3.66 and -3.0910 values between -3.09 and -2.539 values between -2.53 and -1.9720 values between -1.97 and -1.4086 values between -1.40 and -0.84355 values between -0.84 and -0.282249 values between -0.28 and 0.29975 values between 0.29 and 0.85142 values between 0.85 and 1.4217 values between 1.42 and 1.981 values between 1.98 and 2.551 values between 2.55 and 3.112 values between 3.11 and 3.672 values between 3.67 and 4.243 values between 4.24 and 4.802 values between 5.36 and 5.931 values between 9.31 and 9.88offnegativepositiveoften

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. @m
  2. @(
  3. @section
  4. itä
  5. @Table
  6. جموع
  7. Ridley
  8. inois
  9. @",
  10. 女方

Suppresses

  1. 未来
  2. future
  3. future
  4. 未来的
  5. Future
  6. 未來
  7. Future
  8. 将来
  9. _future
  10. futuro