Neuron Explorer

Qwen3-0.6B

"at least" phrase

A neuron in layer 4 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 "least" when it follows "at", as in "at least twice a day" or "at least one". Turning it up did not visibly make the model write about this phrase.

Written by Claude from this neuron's activations and steering outputs. Confidence: high. 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

    , at least. The reason? Google and its ever-changing algorithms. Specifically, one of the

  2. #2 peak

    at least the oft-feared hard landing. The best reply I could give was that it

  3. #3 peak

    at least 5000 BC, when the area still received substantial rainfall, in

  4. #4 peak

    to impact memory and contextual learning, which makes it difficult to distinguish danger versus safety. Thus, brain changes in response to early childhood toxic stress could explain, at least

  5. #5 peak

    clean slate" scenario): (1-2) the two parties in a dialogue, (3) at least one trusted introducer, and (4) at least one trusted witness. Trusted introducers and trusted witnesses allow you to build two open

  6. #6 peak

    sometimes affect the development of the permanent teeth.⏎Don’t forget that your baby needs more than dental visits to have a healthy mouth. Brush your child’s teeth at least twice a day, avoid sugary drinks or juice, don’t use bottles right

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: they are using a lot of different social media platforms. The new student community at the university has started to use their own social media, and the university has been getting more popular than the old. They are using a lot of different platforms,

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

    , we still want to hear from you.⏎1. University Police will assess the situation for danger of imminent harm.⏎2. A team led by the Dean of Students and the Chief Diversity Officer will review the report and reach out to you to provide

  2. #2 peak

    pts/pm/research/)⏎The resources below provide activities, resources, and guidance to supporting the mind, body and heart to cultivate positive well-being.⏎Director of Student Services<|endoftext|>The Rape of a Goddess: How Demeter Beat the All-Powerful

  3. #3 peak

    corridors. Each building has several dust bins from where the housekeeping staff collects the trash. In the same manner waste from canteens, residential quarters, Hostel and guest houses is collected. The collected waste is dumped in big containers (

  4. #4 peak

    ner will share his personal story of survival and reflect on how the horrors of the Holocaust could have happened in a lecture sponsored by Willamette University’s Jewish Student Union.⏎The free event, which will

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. L2:111 weakly suppresses
  2. L2:2345 weakly suppresses
  3. L3:2794 weakly excites
  4. L0:2677 weakly excites
  5. L3:691 weakly excites
  6. L3:294 weakly excites
  7. L3:456 weakly suppresses
  8. L1:2063 weakly excites
  9. L3:2369 weakly suppresses
  10. L1:613 weakly excites
  11. L2:904 weakly excites
  12. L2:2328 weakly suppresses
  13. L3:1954 weakly excites
  14. L2:2021 weakly suppresses
  15. L3:469 weakly suppresses
  16. L0:1442 weakly excites
  17. L0:270 weakly suppresses
  18. L2:493 weakly suppresses
  19. L3:2981 weakly suppresses
  20. L3:867 weakly excites
  21. L2:49 weakly suppresses
  22. L2:2122 weakly suppresses
  23. L0:556 weakly suppresses
  24. L3:2747 weakly excites

Feeds later layers

  1. L8:1100 suppresses
  2. L6:1673 suppresses
  3. L9:1013 excites
  4. L5:354 suppresses
  5. L8:1483 suppresses
  6. L18:1667 suppresses
  7. L5:2699 suppresses academic and professional terms, college
  8. L7:1259 suppresses
  9. L5:2765 weakly suppresses
  10. L22:2967 weakly suppresses
  11. L5:1153 weakly excites 'least' in 'at least'
  12. L22:1794 weakly excites "shown in" before a figure reference
  13. L6:1042 weakly suppresses
  14. L5:1203 weakly excites
  15. L12:2821 weakly excites
  16. L6:1037 weakly excites
  17. L7:2399 weakly excites
  18. L20:967 weakly excites
  19. L10:2098 weakly excites
  20. L5:1686 weakly suppresses the word 'tropical'
  21. L5:627 weakly excites
  22. L21:1120 weakly excites
  23. L5:2100 weakly excites
  24. L20:506 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 -1.26 and -1.201 values between -1.20 and -1.151 values between -1.09 and -1.031 values between -0.75 and -0.691 values between -0.64 and -0.583 values between -0.52 and -0.464 values between -0.46 and -0.4110 values between -0.41 and -0.3523 values between -0.35 and -0.2949 values between -0.29 and -0.24126 values between -0.24 and -0.18203 values between -0.18 and -0.12359 values between -0.12 and -0.07677 values between -0.07 and -0.01851 values between -0.01 and 0.05651 values between 0.05 and 0.10475 values between 0.10 and 0.16256 values between 0.16 and 0.22117 values between 0.22 and 0.2756 values between 0.27 and 0.3323 values between 0.33 and 0.399 values between 0.39 and 0.443 values between 0.44 and 0.501 values between 0.50 and 0.561 values between 0.56 and 0.612 values between 0.61 and 0.671 values between 0.67 and 0.731 values between 0.73 and 0.781 values between 0.90 and 0.951 values between 0.95 and 1.01offnegativepositiveoften

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. ian
  2. çon
  3. IE
  4. icha
  5. ook
  6. ahun
  7. hec
  8. IAN
  9. 篇章
  10. breadcrumbs

Suppresses

  1. rollable
  2. lạ
  3. `(
  4. uates
  5. физ
  6. 后排
  7. `'
  8. -AA
  9. bucks
  10. IFORM