Neuron Explorer

Qwen3-0.6B

mid-word fragments of long words

A neuron in layer 26 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 tends to fire on a piece of a longer word that has been split up, such as 'Ful' in 'Fuligulinae' or 'ker' in 'handkerchief'. Turning it up did not produce an obvious theme; turning it down made the model repeat the word 'moments'.

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

    ustrades were ornately molded in pre-cast faux stone and lighting was generously provided from electric globes on eleborate cast-iron poles… large plinths marked the foot of the structure at both ends. The four rusticated pl

  2. #2 peak

    Dictionary and Cyclopedia⏎- n. A sea-duck of the subfamily Fuligulinœ and genus Clangula; a garrot. The common golden

  3. #3 peak

    54th Union Army regiment, the first all-volunteer black regiment. The sculpture, dedicated to black soldiers, is located in Boston

  4. #4 peak

    ant that contains at least 70% alcohol.⏎- Other recommendations⏎- Avoid keeping your smartphone and handkerchief in the same pocket. Frequently wash your hands, especially when you are in

  5. #5 peak

    early as February-April. Crops that need to be raised in warmth either indoors, in a greenhouse or under cloches e.g. tomatoes, peppers, courgettes are planted out after being ‘hardened-off’ (kept out during

  6. #6 peak

    Platonis et Aristotelis / Comparison of Plato and Aristotle (1458). This book accused the humanist circle patronized by the Greek-born Platonist Cardinal Johannes Bessar

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: moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments moments

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

    angles formed by the cords on the disk—the sum of the forces and the equilibrium conditions. Provenance: Lorraine collections.<|endoftext|>White sauce,

  2. #2 peak

    and limitations cannot be understood only through electrical specifications and nameplate data, because redundancy is included in the design. In order to obtain a thorough understanding of electromagnetic force/torque generation, one must consider thermal characteristics

  3. #3 peak

    of oil and sweat in the skin. It becomes scaly and even wounds and scars take more time to restore to health.⏎Repetitive facial expressions cause muscle contractions. As the elasticity of the skin diminishes,

  4. #4 peak

    submaximal voluntary isometric contractions with the arm in a pronated position for ten healthy and young male subjects

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. L23:2530 suppresses
  2. L20:2641 suppresses
  3. L16:39 suppresses
  4. L22:1754 weakly excites
  5. L20:2922 weakly excites
  6. L10:1486 weakly excites
  7. L24:105 weakly suppresses
  8. L3:3020 weakly excites
  9. L11:2029 weakly excites
  10. L0:2757 weakly excites
  11. L5:2165 weakly suppresses
  12. L20:295 weakly excites
  13. L14:717 weakly excites
  14. L23:115 weakly excites
  15. L10:876 weakly suppresses
  16. L24:1182 weakly suppresses
  17. L5:593 weakly suppresses
  18. L9:2087 weakly suppresses
  19. L6:1400 weakly excites
  20. L22:273 weakly excites
  21. L16:110 weakly suppresses
  22. L18:633 weakly excites
  23. L3:388 weakly suppresses
  24. L0:2578 weakly suppresses

Feeds later layers

  1. L27:422 weakly suppresses
  2. L27:1073 weakly suppresses
  3. L27:1361 weakly excites
  4. L27:1246 weakly excites
  5. L27:2136 weakly suppresses capital M inside abbreviations
  6. L27:1816 weakly suppresses
  7. L27:1326 weakly excites
  8. L27:3036 weakly suppresses
  9. L27:1915 weakly suppresses
  10. L27:303 weakly suppresses
  11. L27:586 weakly suppresses
  12. L27:2012 weakly suppresses
  13. L27:489 weakly suppresses
  14. L27:1201 weakly suppresses
  15. L27:2786 weakly suppresses lowercase 'm' inside split words
  16. L27:44 weakly suppresses
  17. L27:2413 weakly excites
  18. L27:945 weakly suppresses
  19. L27:2138 weakly excites
  20. L27:2114 weakly suppresses
  21. L27:2113 weakly suppresses
  22. L27:1117 weakly suppresses
  23. L27:2070 weakly excites
  24. L27:2544 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 -55.84 and -53.981 values between -18.60 and -16.741 values between -16.74 and -14.881 values between -11.15 and -9.293 values between -9.29 and -7.433 values between -7.43 and -5.5715 values between -5.57 and -3.7124 values between -3.71 and -1.842415 values between -1.84 and 0.021387 values between 0.02 and 1.8824 values between 1.88 and 3.7417 values between 3.74 and 5.614 values between 5.61 and 7.474 values between 7.47 and 9.332 values between 9.33 and 11.191 values between 11.19 and 13.052 values between 14.92 and 16.783 values between 16.78 and 18.64offnegativepositiveoften

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. Caller
  2. .newaxis
  3. edly
  4. vari
  5. 弦
  6. 正宗
  7. дей
  8. ;$
  9. alan
  10. directory

Suppresses

  1. moments
  2. moment
  3. mom
  4. tor
  5. Moments
  6. Mom
  7. moment
  8. Moment
  9. 矩
  10. mom