Neuron Explorer

Qwen3-0.6B

the letter y starting a word

A neuron in layer 26 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 lights up on a lone "y" that starts a word, such as "yawn", "yam" or "yells". Turned down, the model repeated the word "tinker" and similar filler, and turned up it wandered off-topic.

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

    medical history and do a physical examination.⏎Medical history questions may include:⏎- When did the excessive yawning begin?⏎- How many times do you yawn per hour or day?⏎- Is it worse in the morning, after lunch

  2. #2 peak

    aur, taught Achilles how to use this herb.⏎Whether or not that is true, the world will never know, but we do know that healers placed rolled yarrow leaves into the nose of those who were bleeding throughout medieval times. Achilles might

  3. #3 peak

    garden⏎In the first volume of “Deer in My Garden”, beloved local garden guru Carolyn Singer wrote, “’Grandma, why don’t you grow yucky flowers?’ ‘Why would I do that, Marcus?’ I asked, respectful

  4. #4 peak

    shipwrecks from the Byzantine Empire, eight of which are now described in a new report.⏎The shipwrecks were discovered at a site called Yenikapi, in Istanbul, in what was a port of the ancient city

  5. #5 peak

    . There, too, the government, along with billions from Europe and the US, has tried to give growers an incentive to change crops. But alternative crops like yuca or rice sell for much less and are harder to get to market than

  6. #6 peak

    provides. The highly nutritious oca root is also grown commecially in New Zealand where it has become a popular root vegetable and where it is commonly known as yam.⏎In North America and Europe, oca root is relatively unknown and not

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 about t t t t t t t t t t t t t t t t t t t t t t t t t t t t t t t t t t t t t t t t t t t t t t

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

    . So closely entwined are their lives that a number of evolutionary biologists argue that dogs and humans have evolved together.⏎Over the past decade, a loose-knit band of scientists in laboratories in the United States and abroad have worked diligently

  2. #2 peak

    the first park set aside to protect man-made wonders?⏎Mesa Verde National Park, Colo.⏎What is the largest U.S. National Park?⏎Wrangell-St. Elias National Park and Preserve, Alaska (13.2 million acres)⏎

  3. #3 peak

    Gothic, Perpendicular Gothic choir, religious fervour, Roman city of Eboracum, Roman ruins, school visits, secret sign, skilled carvers, St Bosa of York, St William of York, stained glass, stonem

  4. #4 peak

    to retrieve the correct value from the data table above.⏎Have you ever built an app based on Excel?⏎Method 1: The easiest way (VLOOKUP & MATCH)⏎Probably the easiest – or at least the shortest – way of returning values

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. L18:405 excites
  2. L20:612 excites
  3. L21:442 suppresses
  4. L5:779 suppresses
  5. L16:143 suppresses
  6. L9:449 excites
  7. L6:1826 excites
  8. L6:328 weakly suppresses
  9. L17:278 weakly excites 'as' in 'as well as' and 'as much as'
  10. L1:542 weakly suppresses
  11. L23:2833 weakly excites
  12. L18:249 weakly excites
  13. L16:17 weakly excites
  14. L22:2314 weakly excites
  15. L2:1795 weakly excites
  16. L20:174 weakly excites
  17. L2:1699 weakly excites
  18. L18:423 weakly excites
  19. L20:1120 weakly excites
  20. L18:131 weakly excites
  21. L18:168 weakly excites
  22. L16:7 weakly excites
  23. L20:97 weakly excites
  24. L17:378 weakly suppresses

Feeds later layers

  1. L27:654 weakly suppresses
  2. L27:1118 weakly excites
  3. L27:2583 weakly excites
  4. L27:2412 weakly excites
  5. L27:2799 weakly suppresses the letter d inside words
  6. L27:1436 weakly suppresses
  7. L27:1028 weakly excites
  8. L27:893 weakly suppresses
  9. L27:1920 weakly excites
  10. L27:2285 weakly suppresses
  11. L27:1594 weakly suppresses
  12. L27:1800 weakly excites
  13. L27:2932 weakly suppresses capital F starting a word
  14. L27:534 weakly suppresses
  15. L27:558 weakly suppresses
  16. L27:1771 weakly excites
  17. L27:1990 weakly excites
  18. L27:204 weakly excites
  19. L27:1042 weakly suppresses
  20. L27:679 weakly suppresses
  21. L27:1867 weakly excites
  22. L27:1866 weakly excites
  23. L27:1405 weakly suppresses
  24. L27:577 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.

1 values between -30.17 and -28.441 values between -21.49 and -19.761 values between -18.02 and -16.291 values between -14.55 and -12.821 values between -12.82 and -11.081 values between -11.08 and -9.351 values between -5.87 and -4.141 values between -4.14 and -2.40308 values between -2.40 and -0.673550 values between -0.67 and 1.0721 values between 1.07 and 2.807 values between 2.80 and 4.542 values between 4.54 and 6.282 values between 6.28 and 8.014 values between 8.01 and 9.752 values between 9.75 and 11.481 values between 13.22 and 14.951 values between 14.95 and 16.691 values between 20.16 and 21.901 values between 37.51 and 39.25offnegativepositiveoften

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. Replies
  2. 議
  3. emi
  4. òng
  5. 泳
  6. メント
  7. letics
  8. Foot
  9. 一角
  10. ゃ

Suppresses

  1. modx
  2. keycode
  3. velte
  4. Sakura
  5. .INTER
  6. tal
  7. krat
  8. rup
  9. ibil
  10. VX