Neuron Explorer

Qwen3-0.6B

'and' joining verb or noun pairs

A neuron in layer 15 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 weakly on the word 'and' or 'or' when it links two similar words or actions in technical or academic writing, like 'design and based' or 'people and the environment'. Its signals are small, and turning it up or down gave repetitive or unrelated text.

Written by Claude from this neuron's activations and steering outputs. Confidence: low. Think it is wrong? Flag it (opens an email to the maker with the neuron already filled in).

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

    , statistical, and analytical methods to find out the cases of errors made by Thai learners, and analyzes the error causes. After proceeding Chinese auxiliary verb classroom teaching designed and based on the comparative analysis results, we found that the correct ratio of selection of

  2. #2 peak

    becomes more active. This makes both learning and teaching English more interesting and memorable.⏎This theme is aimed at helping the students realize and nurture their talents and skills. The sessions will provide the students the opportunity to discover

  3. #3 peak

    is also a galvanized steel square plate coated with a thin layer of zinc on the other side, namely a double-sided differential galvanized plate.⏎5. Alloy and composite galvanized steel plate. It is an alloy made of

  4. #4 peak

    Rhythm is the focus of pop music, and melody is the focus of classical music. The mind perceives and organizes music in learned, consistent ways in order to generate predictions and extract meaning

  5. #5 peak

    the protection of people and the environment from harmful effects of ionizing radiation and for the safety of radiation sources

  6. #6 peak

    . The study shows that the long-term shear strength of fractured rocks is mainly composed by two mechanisms: one is the anti-shear due to the interlocking

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, as a member of the, in the (or in the, or as a member/ in, or as in the, or in, or, in, or as, in, or, in, or, in, in

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 are posting the Economics– Important Questions Bank for Rajasthan Board HSC 2016 Examination to make life easy for all you HSC students.⏎Without making you wait any further, please find the questions below:⏎SENIOR SECOND

  2. #2 peak

    as early as in 1920’s under the aegis of freedom fighters such as Netaji Subash Chandra Bose and Rajendra Prasad to name

  3. #3 peak

    atomic structure⏎and a definite chemical composition. The basic⏎source of all minerals is the hot magma in the⏎interior of the earth.⏎(more content follows the advertisement below) A D V E R T I S E M E N

  4. #4 peak

    considered. The other common treatment option is surgery, which consists of two different types: open release and endoscopic surgery.⏎(Click Carpal Tunnel Syndrome to learn more, including what causes the condition and who is most at risk for it. You

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. L10:556 excites
  2. L10:370 suppresses
  3. L9:74 excites
  4. L4:92 suppresses
  5. L3:249 excites
  6. L10:1289 suppresses
  7. L10:1278 suppresses
  8. L10:197 excites
  9. L9:281 excites
  10. L11:223 excites
  11. L11:840 suppresses
  12. L3:23 suppresses
  13. L8:328 excites
  14. L9:207 suppresses
  15. L9:2581 excites
  16. L4:76 excites
  17. L10:873 excites
  18. L2:24 suppresses
  19. L7:92 excites
  20. L13:44 suppresses inside web addresses and citations
  21. L10:155 excites
  22. L4:42 suppresses
  23. L4:236 excites
  24. L9:46 suppresses

Feeds later layers

  1. L17:328 suppresses
  2. L17:785 suppresses
  3. L16:2547 weakly excites names and titles inside proper nouns
  4. L23:139 weakly suppresses
  5. L16:2360 weakly excites
  6. L20:587 weakly excites
  7. L23:1072 weakly suppresses
  8. L21:2148 weakly suppresses
  9. L21:2318 weakly suppresses
  10. L17:936 weakly excites
  11. L16:1802 weakly suppresses
  12. L16:2304 weakly excites
  13. L21:514 weakly excites
  14. L19:176 weakly excites
  15. L18:2398 weakly excites
  16. L22:2506 weakly excites
  17. L17:423 weakly suppresses
  18. L17:2472 weakly excites
  19. L16:124 weakly excites
  20. L23:46 weakly excites
  21. L19:2545 weakly excites
  22. L18:2650 weakly excites
  23. L18:2923 weakly excites
  24. L23:2852 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 -2.02 and -1.931 values between -1.84 and -1.751 values between -1.58 and -1.491 values between -1.31 and -1.232 values between -1.14 and -1.052 values between -0.96 and -0.872 values between -0.87 and -0.7911 values between -0.79 and -0.709 values between -0.70 and -0.619 values between -0.61 and -0.5229 values between -0.52 and -0.4447 values between -0.44 and -0.3578 values between -0.35 and -0.26187 values between -0.26 and -0.17417 values between -0.17 and -0.081325 values between -0.08 and 0.011107 values between 0.01 and 0.09377 values between 0.09 and 0.18151 values between 0.18 and 0.2778 values between 0.27 and 0.3631 values between 0.36 and 0.4519 values between 0.45 and 0.5311 values between 0.53 and 0.623 values between 0.62 and 0.712 values between 0.71 and 0.805 values between 0.80 and 0.882 values between 1.41 and 1.50offnegativepositiveoften

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. OSH
  2. ivant
  3. ザー
  4. erm
  5. tram
  6. ignant
  7. tras
  8. 不再是
  9. trak
  10. igr

Suppresses

  1. |
  2. dịp
  3. | ⏎
  4. เขา
  5. |⏎
  6. 揖
  7. ="/">
  8. ']:
  9. داء
  10. ัก