Neuron Explorer

Qwen3-0.6B

circulation, spread and popularity words

A neuron in layer 4 of Qwen3-0.6B. Its strongest snippets are mixed: they peak on different tokens. It is featured, so it has steering outputs.

What this neuron does

This neuron lights up on words about things spreading out to many people, like 'circulation', 'circulated', 'popularity' and 'marketing'. When it was turned up, the model's replies did not clearly change in any particular direction.

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

    The Acts of Thomas is a gnostic work that seems to have been mentioned by Epiphanius of Salamis, thus strongly suggesting that it was in circulation in the fourth century; it likely goes back to early third century Edessa,

  2. #2 peak

    the official magazine of Sallman's denomination. A publishing firm affiliated with the Church of God (Anderson) saw potential in his work, and began mass-marketing them. He created several versions over the years, the last and most popular of

  3. #3 peak

    been performed in the United States since FDA approval in 2001 and in Europe and Australia since the 1970's. Currently increasing in popularity, laparoscopic sleeve gastrectomy surgery, offers patients another option in surgical

  4. #4 peak

    the climate change that has already been set in motion.⏎Enter a study prepared by Oakland’s Pacific Institute for the California Energy Commission, published in July but not widely circulated until this month. It fills in the holes in the city’s approach by advancing

  5. #5 peak

    group. The telling from generation to generation reinforces the group’s identity…⏎What does folk tale mean?⏎: a characteristically anonymous, timeless, and placeless tale circulated orally among a people.⏎Is Cinderella a folktale or fairy tale?⏎

  6. #6 peak

    there is considerable doubt as to the efficacy of that pathway amid a global demand collapse. so the primary intention is to increase the currency in circulation in the hopes that more on hand will mean more spending and (critically)

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 can make a statement that is the most appropriate and the most accurate in terms of the subject of the statement. What should I do? I need to be more specific? Also, I can take the time to learn and grow and I

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

    How often do you shop online?" And the second is, "Are you male or female?"⏎In this

  2. #2 peak

    time.⏎It is then laid down that both texts are equally authentic, carry the same weight and are equally valid. In the same way, ratifications and accessions will be valid for the two versions.⏎The solution thus adopted conforms to the most

  3. #3 peak

    for products purchased in a shopping district with trees. Those same shoppers also say they are willing to stay longer and rate the products and stores as higher quality in a shopping district with the trees.⏎Trees in the landscape relax us, lower heart rates,

  4. #4 peak

    with online financial transactions. However, occasionally something goes wrong.⏎Here are 10 tips on how to protect yourself when you’re banking, paying bills, or shopping online from the Federal Deposit Insurance Corporation:⏎1. If you bank online, frequently

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. L3:322 suppresses
  2. L3:142 suppresses
  3. L3:2513 suppresses
  4. L3:2925 weakly suppresses
  5. L0:2985 weakly suppresses
  6. L3:2292 weakly suppresses
  7. L0:875 weakly excites
  8. L3:1851 weakly excites
  9. L1:2989 weakly excites
  10. L3:563 weakly suppresses
  11. L3:421 weakly suppresses
  12. L3:2792 weakly suppresses
  13. L2:1098 weakly suppresses
  14. L3:1829 weakly excites
  15. L3:1295 weakly excites
  16. L0:2960 weakly suppresses
  17. L0:2072 weakly excites
  18. L3:294 weakly excites
  19. L2:2785 weakly suppresses
  20. L3:50 weakly excites
  21. L3:1021 weakly excites
  22. L3:3013 weakly suppresses
  23. L3:1894 weakly suppresses
  24. L3:2370 weakly excites

Feeds later layers

  1. L5:2110 excites
  2. L11:2919 weakly excites
  3. L20:2803 weakly excites
  4. L16:1673 weakly excites
  5. L13:1438 weakly suppresses
  6. L20:2559 weakly excites
  7. L7:2793 weakly suppresses
  8. L5:2684 weakly suppresses
  9. L5:2577 weakly excites
  10. L19:2979 weakly excites
  11. L6:2216 weakly excites
  12. L6:2809 weakly excites
  13. L11:278 weakly suppresses
  14. L18:2146 weakly excites
  15. L13:1763 weakly excites 'know' after 'did you' and how-to verbs
  16. L7:1028 weakly excites
  17. L24:2702 weakly excites
  18. L20:2420 weakly excites
  19. L9:2219 weakly excites
  20. L21:3027 weakly excites
  21. L5:832 weakly suppresses
  22. L6:367 weakly excites
  23. L9:1030 weakly excites
  24. L21:2462 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.

1 values between -0.95 and -0.813 values between -0.81 and -0.665 values between -0.66 and -0.529 values between -0.52 and -0.37105 values between -0.37 and -0.23784 values between -0.23 and -0.082066 values between -0.08 and 0.06753 values between 0.06 and 0.21134 values between 0.21 and 0.3522 values between 0.35 and 0.507 values between 0.50 and 0.644 values between 0.64 and 0.793 values between 0.79 and 0.932 values between 0.93 and 1.083 values between 1.08 and 1.221 values between 1.37 and 1.511 values between 1.51 and 1.651 values between 2.23 and 2.381 values between 2.96 and 3.102 values between 3.82 and 3.971 values between 4.69 and 4.84offnegativepositiveoften

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. hlen
  2. *@
  3. 高价
  4. ثان
  5. 主义思想
  6. 纸上
  7. 出去
  8. /share
  9. itm
  10. 嗅

Suppresses

  1. icare
  2. cares
  3. iculo
  4. ≻
  5. ercul
  6. care
  7. letic
  8. 园区
  9. /al
  10. ulumi