Neuron Explorer

Qwen3-0.6B

unclear: mixed word fragments

A neuron in layer 25 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

Nothing clear ties this neuron together. It fires on a grab bag of word pieces in very different topics, such as a Dutch-Afrikaans name, a geology term, an engineering term, and a welding term. Turning it up made the model ramble in broken, repetitive text.

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

    Voortrekkers comprised two groups from the eastern frontier region of the Cape Colony, semi

  2. #2 peak

    is obtained from drainage observations and a surface correlation of the maps.⏎All five caves occur in the Tomstown Dolomite (dip 25º E., strike N.-S.). The area itself is almost entirely surrounded by metamorphic or ign

  3. #3 peak

    is effective in controlling mosquitoes that transmit human diseases such as Aedes spp., Anopheles gambiae, Anopheles culicifacies, Bamisia tabaci, Cule<|endoftext|>The American Association of Feline Practitioners (AA

  4. #4 peak

    687|⏎|Julkaisu||JOURNAL OF LANGUAGE TEACHING AND RESEARCH|⏎|DOI - pysyväislinkit|⏎|Tila||Julkaistu - 1 syysk. 20

  5. #5 peak

    directional or reverse current relays differential relays, and distance relays.⏎According to timing characteristics the relays can be divided into instantaneous, definite time lag, inverse time lag and inverse definite minimum time (IDMT) lag relays.⏎In

  6. #6 peak

    where four major flyways converge. This area is also along a major migration route bordering the Gulf of Mexico, where many common species that winter in the Neotropics might linger until Christmas.⏎For more information on the annual count, visit

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: one who who is who who who who who who who who who who who who who who who who who who who who who who who who who who who who who who who who who who who who who who who who who who who who

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

    data points, it mainly deals with unlabelled data.⏎Some applications of unsupervised machine learning techniques include: clustering (grouping coherent segments of customers,

  2. #2 peak

    , we enhance these assumptions by considering mechanically diverse layers with varying dips. We employed the Finite Element Method (FEM) to create numerical models simulating two chamber shapes: a circular form and

  3. #3 peak

    shipbuilding, aerospace, prosthetics, architectural projects and even special effects in movies. It is essentially the evolution of a merger between a graphical user interface and Non-Uniform Rational B-Splines (NURBS) geometry, and currently functions

  4. #4 peak

    designs show significant improvement over microelectronic designs in metrics such as footprint area and power consumption. These improvements are obtained from designs employing nanodevices and nanofabrication techniques that already have been demonstrated experimentally. Thus, the results presented

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. L20:238 excites
  2. L23:115 excites
  3. L10:656 excites
  4. L4:879 weakly excites
  5. L3:105 weakly excites
  6. L3:838 weakly suppresses
  7. L3:1333 weakly excites
  8. L19:746 weakly excites
  9. L16:974 weakly suppresses
  10. L7:1328 weakly suppresses
  11. L20:139 weakly excites
  12. L11:3046 weakly excites
  13. L2:2447 weakly excites
  14. L10:763 weakly suppresses
  15. L9:1917 weakly suppresses
  16. L16:270 weakly excites
  17. L6:1991 weakly excites
  18. L7:1071 weakly suppresses
  19. L12:3041 weakly excites
  20. L5:78 weakly excites
  21. L12:69 weakly excites
  22. L1:2612 weakly suppresses
  23. L18:2681 weakly excites
  24. L8:2721 weakly suppresses

Feeds later layers

  1. L27:2511 weakly excites
  2. L27:541 weakly suppresses
  3. L26:2524 weakly excites
  4. L26:3024 weakly excites
  5. L27:2266 weakly suppresses
  6. L27:2529 weakly suppresses
  7. L27:1335 weakly excites
  8. L26:1224 weakly suppresses
  9. L26:2894 weakly excites
  10. L27:2893 weakly excites
  11. L27:1726 weakly excites
  12. L27:4 weakly excites
  13. L26:754 weakly excites
  14. L26:2023 weakly suppresses
  15. L26:444 weakly excites
  16. L27:2392 weakly excites
  17. L26:3041 weakly excites
  18. L27:1991 weakly excites
  19. L26:462 weakly excites
  20. L27:1051 weakly excites
  21. L26:1722 weakly excites
  22. L26:2938 weakly excites
  23. L26:1550 weakly suppresses
  24. L27:2010 weakly excites

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 -10.70 and -10.101 values between -10.10 and -9.502 values between -9.50 and -8.891 values between -8.89 and -8.291 values between -8.29 and -7.694 values between -7.69 and -7.095 values between -7.09 and -6.492 values between -6.49 and -5.896 values between -5.89 and -5.291 values between -5.29 and -4.696 values between -4.69 and -4.0910 values between -4.09 and -3.4911 values between -3.49 and -2.8919 values between -2.89 and -2.2933 values between -2.29 and -1.6966 values between -1.69 and -1.09270 values between -1.09 and -0.491824 values between -0.49 and 0.111309 values between 0.11 and 0.71210 values between 0.71 and 1.3141 values between 1.31 and 1.9129 values between 1.91 and 2.5119 values between 2.51 and 3.1111 values between 3.11 and 3.716 values between 3.71 and 4.317 values between 4.31 and 4.924 values between 4.92 and 5.522 values between 5.52 and 6.121 values between 6.12 and 6.721 values between 7.32 and 7.922 values between 7.92 and 8.521 values between 9.12 and 9.721 values between 9.72 and 10.321 values between 12.72 and 13.32offnegativepositiveoften

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. ete
  2. etes
  3. الداخلية
  4. ValueEventListener
  5. itempty
  6. ewis
  7. (grammarAccess
  8. 子弟
  9. öße
  10. akov

Suppresses

  1. .gg
  2. 죽
  3. อง
  4. IGN
  5. alty
  6. 模式
  7. yr
  8. 名义
  9. richt
  10. -channel