Neuron Explorer

Qwen3-0.6B

'for' as in 'for example'

A neuron in layer 23 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 'for' when it sits in an aside like 'if, for example' or 'for some reason'. Turning it up or down made the model fall into repeated filler words like 'the the the' or 'from from', so the steering results are not very readable.

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

    All U.S. states have neonatal screening for CAH. Infants who test positive need to have follow-up testing done to confirm the diagnosis. If, for some reason, the neonatal screening is negative but there is high suspicion for CA

  2. #2 peak

    ⏎How do people determine lengths when they can't use standard measuring tools, such as a tape measure? How do they find the distance across a river if, for example, they don't have a boat? Measurements of very large objects or of

  3. #3 peak

    US, and Environment Canada - WILDSPACE.⏎- Migration/Movement:Elevational Migrant⏎- Primary Habitat:Montane evergreen forest edge⏎- Foraging Strata:Understory/Canopy⏎- Foraging Behavior:Probe

  4. #4 peak

    molecules into long, straight rods lets electrons flow freely, approximating the conductivity of traditional materials like silicon or copper. Heeger and his co-Nobelists, for example, discovered that oxidizing the polymer polyacetylene with iodine vap

  5. #5 peak

    to them.⏎The exhibit tells the story of migrant farmworkers, their work, and the people who serve them, in a number of ways. Photographs, for the most part by the superbly skilled professional photographers who have been acknowledged previously,

  6. #6 peak

    Inventory Number: AUT 040<|endoftext|>Communities in Kenya have adopted one another’s food preferences and recipes. With the exception of the Coast and Indian communities for whom food preparation is an elaborate process, nearly all other communities garnish their food

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: from of of of from of from of from of from of from of from of from of from of from of from of from of from of from of from of from of from of from of from of from of from of from of from of

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

    get can relies on the amount ingested. My pet has had tiny amounts of onion and garlic and it hasn't caused any issues. He once additionally stole an After-Eight! Mint, once more - this was simply a tiny low quantity of

  2. #2 peak

    following lesson was created by Vimeo for its Vimeo Video School . It’s used here with permission. Look for a new lesson each week.⏎Want to get started with After Effects but don’t know where to begin? In this lesson Nick Campbell will guide

  3. #3 peak

    playback the piece you’re working on⏎- Create a pre-comp to easily duplicate certain animations for re-use⏎Here’s another great example video done by Nick in After Effects using precomps!⏎Check back next week to learn the difference between frame

  4. #4 peak

    and consultations with the general public and initiated public campaigns to encourage people to voluntarily turn in their firearms. The government’s new security initiative “Gran Misión a Toda Vida Venezuela” pledged to continue this work of disarmament, including through the

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. L20:593 excites
  2. L20:283 excites
  3. L18:293 suppresses
  4. L17:303 suppresses
  5. L22:353 suppresses
  6. L17:1223 excites
  7. L18:2153 suppresses
  8. L4:4 suppresses
  9. L13:132 excites "of" after a noun
  10. L20:64 suppresses
  11. L6:1534 suppresses
  12. L8:45 suppresses
  13. L10:132 excites
  14. L19:1980 excites
  15. L17:706 excites
  16. L6:388 excites
  17. L19:2449 suppresses
  18. L17:385 suppresses
  19. L10:272 suppresses
  20. L12:234 excites
  21. L17:317 excites 'of' after a noun phrase
  22. L17:264 suppresses
  23. L22:554 suppresses
  24. L7:8 excites

Feeds later layers

  1. L25:247 excites
  2. L27:149 suppresses
  3. L27:474 excites
  4. L27:108 suppresses
  5. L27:121 suppresses
  6. L24:49 suppresses
  7. L26:693 excites
  8. L27:663 excites
  9. L26:71 suppresses
  10. L26:139 suppresses
  11. L24:92 suppresses
  12. L24:116 excites
  13. L27:262 excites
  14. L27:168 excites
  15. L26:631 weakly excites
  16. L25:802 weakly suppresses
  17. L24:604 weakly suppresses
  18. L27:528 weakly excites
  19. L25:1220 weakly suppresses
  20. L26:920 weakly suppresses
  21. L27:244 weakly excites
  22. L25:306 weakly suppresses
  23. L27:810 weakly excites
  24. L26:964 weakly excites "in" after a publication or study word

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.

3 values between -7.35 and -5.7311 values between -5.73 and -4.129 values between -4.12 and -2.50270 values between -2.50 and -0.882823 values between -0.88 and 0.74288 values between 0.74 and 2.3586 values between 2.35 and 3.9766 values between 3.97 and 5.5951 values between 5.59 and 7.2141 values between 7.21 and 8.8226 values between 8.82 and 10.4423 values between 10.44 and 12.0627 values between 12.06 and 13.6727 values between 13.67 and 15.2921 values between 15.29 and 16.9117 values between 16.91 and 18.539 values between 18.53 and 20.149 values between 20.14 and 21.7610 values between 21.76 and 23.3812 values between 23.38 and 25.009 values between 25.00 and 26.615 values between 26.61 and 28.236 values between 28.23 and 29.859 values between 29.85 and 31.477 values between 31.47 and 33.085 values between 33.08 and 34.704 values between 34.70 and 36.322 values between 36.32 and 37.946 values between 37.94 and 39.556 values between 39.55 and 41.172 values between 41.17 and 42.794 values between 42.79 and 44.411 values between 44.41 and 46.024 values between 46.02 and 47.641 values between 47.64 and 49.261 values between 50.87 and 52.493 values between 52.49 and 54.112 values between 54.11 and 55.732 values between 55.73 and 57.34offnegativepositiveoften

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. some
  2. this
  3. 一些
  4. the
  5. various
  6. several
  7. any
  8. those
  9. many
  10. one

Suppresses

  1. iously
  2. onian
  3. 也不敢
  4. 因为他
  5. 网首页
  6. from
  7. 留给
  8. 因为
  9. 说出来
  10. 眨眼