Neuron Explorer

Qwen3-0.6B

"the" after "over" in time spans

A neuron in layer 20 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 "the" in phrases like "over the past decade" or "over the years". Turned up, the model produced strings of "the first" and "last" words.

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

    existing finish rather than create a new one in the restoration process. This process may risk the loss of the remaining<|endoftext|>With the development of computers and the internet over the past several decades, communication and written language have undergone drastic transformations. Until a couple

  2. #2 peak

    date backwards, using B.C. dates, to the time of the creation of the world. Provisionally, based on the studies published in Biblical Chronology over the past two years, I have come to the year 3932 B

  3. #3 peak

    colonies are disappearing fast. According to the nonprofit organization Pollinator Partnership, the United States has lost over 50 percent of its managed honeybee colonies over the past 10 years.⏎Pollinator gardens are just one of the many ways

  4. #4 peak

    . “The shellfish results indicate the strength of the growing aquaculture sector.”⏎Supply and demand of farmed oysters have risen exponentially in North Carolina over the past

  5. #5 peak

    proximity to our brains, many people are worried that they might be associated with an increased risk of brain tumors. There have been around 30 research studies over the past decade that have looked at this issue and most have found that:⏎- People

  6. #6 peak

    of large balloons designed to beam the Internet to unwired places—developed the technology over the last 18 months. Google’s blog reports:⏎“Over the years, many scientists have investigated various body fluids—such as tears—in the hopes

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: I am a 36-year-old, 43-bmd, 56-bmd, 56-bmd, 44-bmd, 46-bmd, 46-bmd, 44

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

    . Does not equal residence. Can only have one domicile at a time.⏎ii. Person starts out life with domicile of origin:⏎1. determined by parents.⏎iii. Acquire a domicile of choice when abandon the domicile of origin; the domicile

  2. #2 peak

    of methane emissions. According to the Environmental Protection Agency, methane can remain in the atmosphere for up to 15 years and is 20 times more likely to trap heat

  3. #3 peak

    Know the Difference⏎- Excessive Heat Watch⏎– Conditions are favorable for an excessive heat event to meet or exceed local Excessive Heat Warning criteria in the next 24 to 72 hours.⏎- Excessive Heat Warning⏎– Heat

  4. #4 peak

    2x√(1+x6)⏎Comment. There is no need to integrate and then differentiate by x.⏎If you want to type x6 instead of x^6 simply type x6, highlight only 6 and click on the icon (

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. L19:1263 excites
  2. L19:3065 excites 'for' after a span of time
  3. L16:1027 excites
  4. L17:1912 suppresses
  5. L17:2544 suppresses
  6. L19:254 suppresses
  7. L5:2700 excites
  8. L19:2789 suppresses
  9. L15:1825 excites
  10. L18:291 suppresses
  11. L18:1487 weakly suppresses
  12. L17:1842 weakly suppresses
  13. L13:2164 weakly excites
  14. L19:2563 weakly excites
  15. L9:1575 weakly excites
  16. L10:2995 weakly excites
  17. L19:1419 weakly excites
  18. L19:2151 weakly excites
  19. L18:3013 weakly suppresses
  20. L18:2750 weakly suppresses
  21. L17:2446 weakly suppresses
  22. L16:1898 weakly excites
  23. L17:2887 weakly excites
  24. L17:2908 weakly excites

Feeds later layers

  1. L21:807 excites
  2. L26:2398 excites
  3. L27:910 excites
  4. L25:869 excites
  5. L24:1775 excites
  6. L22:1470 excites
  7. L26:1094 excites
  8. L23:892 excites "the" in "over the past" time spans
  9. L22:3024 excites
  10. L21:1797 excites
  11. L23:1339 excites
  12. L23:982 excites
  13. L21:1248 excites
  14. L23:3063 excites
  15. L24:2546 excites
  16. L24:2353 weakly excites
  17. L24:549 weakly excites
  18. L21:2207 weakly excites
  19. L22:2987 weakly excites
  20. L21:2283 weakly excites
  21. L21:2559 weakly excites
  22. L24:995 weakly excites
  23. L24:1835 weakly excites
  24. L22:2156 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.

67 values between -1.97 and -1.20506 values between -1.20 and -0.433076 values between -0.43 and 0.34206 values between 0.34 and 1.1013 values between 1.10 and 1.875 values between 1.87 and 2.645 values between 2.64 and 3.414 values between 3.41 and 4.183 values between 4.18 and 4.951 values between 4.95 and 5.726 values between 5.72 and 6.492 values between 6.49 and 7.252 values between 7.25 and 8.022 values between 8.02 and 8.791 values between 9.56 and 10.331 values between 10.33 and 11.102 values between 11.87 and 12.642 values between 12.64 and 13.411 values between 13.41 and 14.171 values between 17.25 and 18.021 values between 18.02 and 18.791 values between 28.01 and 28.78offnegativepositiveoften

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. last
  2. last
  3. first
  4. 最后一个
  5. Last
  6. first
  7. (last
  8. [last
  9. .last
  10. 第一个

Suppresses

  1. yours
  2. 可能です
  3. 属实
  4. onden
  5. ours
  6. ่าง
  7. ancement
  8. itud
  9. RoutingModule
  10. 습