Neuron Explorer

Qwen3-0.6B

the digit 1 in numbers

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 digit '1' when it starts a number, such as the first digit of 140/90 or 1230 degrees. It is a number-position detector, and the steering results do not show a clear effect of turning it up or down.

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

    This can be caused by atherosclerosis or narrowing of the blood flow in the heart. ...Read moreSee 1 more doctor answer⏎My BP is 140/120 sometimes, other times 130/9

  2. #2 peak

    Porcelain is rolled to create the tile, each letter is hand imprinted into the clay, which is then fired in the potter’s kiln at 1230 degrees celsius, making it solid to withstand all elements. The

  3. #3 peak

    finally evaporating at 212 degrees, making the cookie stiffer, but leaving behind empty spaces that makes it lighter and flaky.⏎- At 144 degrees, the proteins in eggs—which look like tightly coiled springs—

  4. #4 peak

    .⏎If your blood pressure reading is high, you will probably be asked to come back for repeat blood pressure checks. If you have three visits with readings over 140/90 mmHG, you will be diagnosed with high blood pressure

  5. #5 peak

    1185 CE), Japanese art flourished, and new forms of art, such as screen painting, began to emerge. The Kamakura period (1185 CE to 1333 CE) saw the rise of

  6. #6 peak

    beginning of the research, 15 subjects were randomly assigned to either a moist-heat or stretching treatment. The moist-heat therapy involved placing a hot (160-degree Fahrenheit), moist pack in a terry-cloth cover on each

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 if the first two degrees of pi, pi plus 3, pi plus 6, pi plus 9, pi plus 12, pi plus 15, and pi plus 18, plus the degrees of pi

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

    standards will be piloted in Botswana, Egypt, Ethiopia, Kenya and Paraguay. The initiative contributes to WHO’s 13th General Program of Work target of ‘1 billion lives made healthier’ by

  2. #2 peak

    Wall Street” protests were against the far-reaching negative consequences of neo-liberal capitalism.⏎In the wake of the global financial meltdown, the notion of Capitalism 4.0 has suddenly made headlines and the Davos Forum early this year also addressed

  3. #3 peak

    – but the story remains. Products of short-lived radioactive isotopes found in some meteorites contradict the 4.5 billion year story. The elements that would have formed primordially in

  4. #4 peak

    Lorenz developed the concept that tiny effects in one part of a system can lead to big changes in another part. The term butterfly effect first appeared in his 1972 paper Predictability: Does the flap of a butterfly's wings in

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. L19:418 excites
  2. L19:723 excites
  3. L19:2533 excites
  4. L16:1333 excites 'two' after 'divided into' or 'there are'
  5. L9:1348 suppresses
  6. L19:720 excites
  7. L17:512 suppresses
  8. L4:205 excites the digit 2 inside citations
  9. L21:1863 suppresses
  10. L19:851 weakly excites comma inside numbers with thousands
  11. L21:2842 weakly excites
  12. L4:464 weakly excites
  13. L19:1047 weakly excites
  14. L16:1012 weakly excites 'one' before 'of' in 'is one of the'
  15. L19:25 weakly excites
  16. L12:953 weakly excites
  17. L21:518 weakly excites
  18. L9:13 weakly excites
  19. L0:17 weakly excites
  20. L9:1032 weakly suppresses
  21. L17:157 weakly excites
  22. L5:2867 weakly excites
  23. L20:416 weakly excites the "1" in years 191x
  24. L3:659 weakly suppresses

Feeds later layers

  1. L24:2097 weakly suppresses
  2. L27:1527 weakly excites
  3. L24:23 weakly suppresses
  4. L23:2634 weakly suppresses
  5. L23:2028 weakly excites
  6. L24:2720 weakly excites
  7. L23:2790 weakly excites
  8. L23:2351 weakly excites
  9. L23:735 weakly excites
  10. L25:2991 weakly excites
  11. L24:1750 weakly excites
  12. L23:1109 weakly excites
  13. L25:786 weakly excites
  14. L25:2908 weakly suppresses
  15. L24:2632 weakly excites
  16. L23:284 weakly excites
  17. L23:2505 weakly excites
  18. L27:1598 weakly excites
  19. L23:1563 weakly excites
  20. L24:2384 weakly excites
  21. L23:599 weakly excites
  22. L25:1278 weakly excites
  23. L25:457 weakly suppresses
  24. L25:2737 weakly excites 'a' before 'few' or 'couple'

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 -22.88 and -21.202 values between -16.17 and -14.491 values between -9.46 and -7.783 values between -7.78 and -6.105 values between -6.10 and -4.426 values between -4.42 and -2.7552 values between -2.75 and -1.073701 values between -1.07 and 0.6165 values between 0.61 and 2.2920 values between 2.29 and 3.968 values between 3.96 and 5.6411 values between 5.64 and 7.325 values between 7.32 and 9.003 values between 9.00 and 10.676 values between 10.67 and 12.354 values between 12.35 and 14.033 values between 14.03 and 15.701 values between 15.70 and 17.381 values between 17.38 and 19.063 values between 19.06 and 20.741 values between 20.74 and 22.411 values between 22.41 and 24.091 values between 24.09 and 25.772 values between 27.45 and 29.121 values between 40.86 and 42.541 values between 42.54 and 44.22offnegativepositiveoften

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. slashes
  2. uries
  3. trusts
  4. UIStoryboardSegue
  5. ifetime
  6. chron
  7. anchor
  8. 阄
  9. 邗
  10. Disqus

Suppresses

  1. eld
  2. 馅
  3. ffield
  4. °C
  5. .CommandType
  6. iative
  7. ICLES
  8. ices
  9. ��
  10. elden