Neuron Explorer

Qwen3-0.6B

the word 'in' after a verb or noun

A neuron in layer 21 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 the small word 'in' when it starts a phrase that follows a verb or noun, like 'in order to' or 'in common with'. Turning it up made the model's text drift into unrelated topics such as grids, numbers and weather, so the link to the word 'in' is the only clear part.

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

    classroom in a setting where theory meets praxis. This course engages students in service-learning, which allows students to contribute their time and energy, in the context of academic coursework that involves reflection and intentional learning goals, in order to

  2. #2 peak

    sandstone.⏎[SinglePic not found]Every year thousands of local and international tourists visit the Phnom Da, Takeo. So many tourists visit the place in order to do a research on the well-known work of art of Cambodian

  3. #3 peak

    osis, I keep going round in circles!⏎Ok, can somebody please answer the following questions:⏎1. In me

  4. #4 peak

    being re-written. Until the TV movie based on Wells' novel was broadcast a few months later, this claim stood as being true. But once "The First Men In The Moon" aired, then Professor Cavor had to be credited with being

  5. #5 peak

    itself or in its effect”.⏎The constitutional anti-discrimination clause is set out in Article 23(2) that reads “no person shall be treated in a discriminatory manner by any person acting by virtue of any written law or in the

  6. #6 peak

    at the end of a person’s full name, after any given names.⏎What exactly is a surname?⏎Also called last name . the name that a person has in common with other

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 have a 50% probability of not having a family n't any family n't any family n't any family n't any family n't any family n't any... n't any... n't any... n't any

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

    are the primary kinds of antique lamps most favored by collectors. In the late 18th century, Swiss chemist Aime Argand invented the first lamp didn't require a free

  2. #2 peak

    the dark and the white races. We find the same differences in proportion between blondes and brunettes in the white races.⏎HAVELOCK ELLIS ON BLONDES AND BRUNETTES⏎The noted anthropologist Havelock Ellis says,

  3. #3 peak

    flak, flack, blast (intense adverse criticism) "Clinton directed his fire at the Republican Party"; "the government has come under attack"; "don't give me any flak"⏎- S: (v) open fire,

  4. #4 peak

    . Celery chipmunks and squirrels endangered, vulnerable, and threatened species flower gardening african wild dogs animal life zoos cats (felines) middle ages donkeys and mules papaya ducks domestic. This makes celery a good and healthy

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. L19:595 strongly excites
  2. L18:282 strongly excites
  3. L3:410 strongly suppresses
  4. L16:55 strongly excites
  5. L10:1745 strongly excites
  6. L1:27 strongly excites
  7. L0:210 strongly excites
  8. L19:231 excites
  9. L17:391 excites
  10. L2:1923 suppresses
  11. L18:1279 excites
  12. L20:1040 excites
  13. L5:499 excites
  14. L0:24 suppresses
  15. L17:725 excites
  16. L14:95 excites
  17. L20:1659 suppresses
  18. L15:115 excites
  19. L15:96 suppresses
  20. L0:781 suppresses
  21. L20:2203 excites
  22. L20:435 excites
  23. L19:1639 suppresses
  24. L19:359 excites

Feeds later layers

  1. L26:2511 strongly suppresses
  2. L22:2299 strongly suppresses
  3. L23:1532 strongly suppresses
  4. L23:2762 strongly suppresses
  5. L27:3060 suppresses 'n' inside word fragments
  6. L24:2458 suppresses
  7. L22:2029 suppresses
  8. L25:2713 suppresses
  9. L25:890 suppresses
  10. L23:3041 suppresses
  11. L24:2929 suppresses
  12. L26:548 excites
  13. L27:2933 suppresses the letter n opening a word
  14. L27:2141 suppresses
  15. L26:2533 suppresses
  16. L25:2112 suppresses
  17. L22:1837 excites
  18. L27:2954 suppresses
  19. L23:2487 suppresses
  20. L24:2480 suppresses
  21. L23:1489 weakly excites
  22. L26:1408 weakly suppresses the 'n' that starts nectar
  23. L24:3047 weakly suppresses
  24. L27:2849 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.

2 values between -56.41 and -54.871 values between -54.87 and -53.342 values between -53.34 and -51.812 values between -51.81 and -50.281 values between -48.75 and -47.221 values between -42.62 and -41.093 values between -41.09 and -39.561 values between -34.96 and -33.431 values between -33.43 and -31.902 values between -31.90 and -30.361 values between -30.36 and -28.833 values between -28.83 and -27.303 values between -27.30 and -25.772 values between -25.77 and -24.241 values between -24.24 and -22.703 values between -22.70 and -21.173 values between -21.17 and -19.641 values between -19.64 and -18.112 values between -16.58 and -15.042 values between -15.04 and -13.512 values between -13.51 and -11.982 values between -11.98 and -10.455 values between -10.45 and -8.926 values between -8.92 and -7.3810 values between -7.38 and -5.8512 values between -5.85 and -4.3220 values between -4.32 and -2.7928 values between -2.79 and -1.261551 values between -1.26 and 0.282175 values between 0.28 and 1.8148 values between 1.81 and 3.3412 values between 3.34 and 4.87offnegativepositiveoften

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. din
  2. d
  3. tas
  4. eenth
  5. ennes
  6. tti
  7. irst
  8. tok
  9. een
  10. tif

Suppresses

  1. ners
  2. ning
  3. NING
  4. etwork
  5. ny
  6. nable
  7. ned
  8. ningen
  9. nen
  10. ights