Neuron Explorer

Qwen3-0.6B

first-person 'I' starting a clause

A neuron in layer 25 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 word 'I' when it begins a clause, such as after 'but' or 'where' ('where I live', 'but I have never'). Turning it up made the model drift into loops of words like 'guess', and turning it down made it repeat 'are'.

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

    many of them residential school survivors, their friends and family.⏎“When I take a look back at all those struggles, I am so pleased that I’m sitting where I am, because it’s not just a random seat being the chancellor of this university

  2. #2 peak

    ito ayuda" "I need help." Or, a fluffly feline sits next to a glass of water, representing "tengo sed," or "I am thirsty."⏎The app also challenges its students to answer increasingly difficult typing, match

  3. #3 peak

    minutes to play, and despite its sophisticated 3D visuals—the game<|endoftext|>Disk or partition alignment is something new to me. I have heard of it before but I have never used nor understood it’s meaning until a colleague at work brought it

  4. #4 peak

    out of the kindness of my heart, because they have some historical significance (as in the case of my review of the LM741) or most likely I think the technology is important and interesting. All opinions are my own and I would

  5. #5 peak

    . The information in this post is taken from a blog posting entitled […]⏎Disk or partition alignment is something new to me. I have heard of it before but I have never used nor understood it’s meaning until a colleague at work brought it to

  6. #6 peak

    Along the way, we have purposely sought out resources and opportunities to develop fluency in a second language.⏎Finding materials for Norwegian is not easy (at least where I live) so I was very excited to discover the New Amigos board game.

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 are high-arching my high-arched high-arched high-arched high-arched high-arched high-arched high-arched high-arched high-arched high-arched high

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

    willow oaks tend to be the most susceptible. Despite the name, these mites can also be found on birch, beech, elm, and hickory trees.⏎Treatment For<|endoftext|>If you have been though the San Diego Zoo

  2. #2 peak

    and called “Break!”—a single word that told McConnell that he was in immediate danger and to take evasive action. But the call came to late. Fedorets fired at McConnell with his MiG-15’s three autocannon

  3. #3 peak

    cause the brain to churn out natural pain-killing endorphins.⏎The new findings add to the scientific heft underlying acupuncture, said neuroscientist Maiken Nedergaard, M.D., D.M.Sc., who led the research.

  4. #4 peak

    growing poorly in sandy or rocky soils. In the wild, walnut is frequently found growing with other tree species including elm, sugar maple, ash, red oak and hickory.⏎Black Walnut Uses⏎High-quality walnut is valued primarily for its use

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. L23:1521 strongly suppresses
  2. L21:951 strongly excites
  3. L20:2578 strongly excites
  4. L19:2758 strongly excites
  5. L24:464 excites
  6. L20:1058 suppresses
  7. L18:2165 excites
  8. L19:2043 excites
  9. L7:67 excites
  10. L18:205 excites
  11. L6:483 suppresses
  12. L5:52 suppresses
  13. L1:1063 excites
  14. L17:2341 excites
  15. L4:1855 excites
  16. L20:2971 suppresses
  17. L21:1040 excites
  18. L23:921 excites
  19. L8:268 excites first person 'I' or 'my' in speech
  20. L22:892 excites
  21. L16:3065 weakly suppresses
  22. L10:1641 weakly suppresses
  23. L7:169 weakly suppresses
  24. L0:2262 weakly excites

Feeds later layers

  1. L27:650 strongly suppresses
  2. L26:72 strongly suppresses 'they' or 'we' after where or who
  3. L26:2947 strongly suppresses
  4. L27:1256 excites
  5. L26:1626 excites the word 'I' as subject
  6. L26:93 excites
  7. L26:1417 suppresses
  8. L27:804 suppresses
  9. L27:422 excites
  10. L26:1235 suppresses opening parenthesis before an acronym
  11. L26:1372 weakly suppresses
  12. L26:1724 weakly excites
  13. L27:746 weakly suppresses
  14. L27:3019 weakly suppresses
  15. L26:418 weakly suppresses
  16. L27:1729 weakly suppresses
  17. L26:874 weakly suppresses
  18. L26:2147 weakly suppresses
  19. L27:2102 weakly suppresses
  20. L26:1753 weakly suppresses
  21. L27:2480 weakly suppresses
  22. L27:1040 weakly suppresses
  23. L27:2324 weakly suppresses
  24. L26:109 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.

2 values between -16.97 and -13.802 values between -13.80 and -10.624 values between -10.62 and -7.459 values between -7.45 and -4.28243 values between -4.28 and -1.113609 values between -1.11 and 2.0724 values between 2.07 and 5.249 values between 5.24 and 8.412 values between 8.41 and 11.591 values between 11.59 and 14.761 values between 14.76 and 17.931 values between 30.62 and 33.791 values between 106.77 and 109.94offnegativepositiveoften

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. am
  2. AM
  3. Am
  4. AM
  5. Am
  6. was
  7. 'am
  8. _am
  9. am
  10. 'm

Suppresses

  1. are
  2. _are
  3. 都是
  4. are
  5. 都
  6. Are
  7. 都不是
  8. 都在
  9. Are
  10. 都是非常