Neuron Explorer

Qwen3-0.6B

body, skin and health words

A neuron in layer 9 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 words about the body and health, such as moisture, dryness, weight, and skin sensitivity. Its strongest negative values come from odd name fragments in the middle of a word, and steering it up did not produce a clear health theme.

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

    • Houses built on a slope are often an issue. They need to be tied back well with footings. Also make sure the slope is stable. Liquefaction -- where saturated soil becomes liquid -- can be a problem and can occur when a

  2. #2 peak

    skin loses its elasticity and onset of sagging and loosening is been visible. One can perceive the formations of crease and the skin loses its capacity to absorb moisture and retain it within the cells. Inadequacy of moisture directly influences the

  3. #3 peak

    factors such as cosmetics, detergents, weather conditions, and even certain foods. People with sensitive face may experience symptoms such as redness, itching, dryness, and inflammation, which can be uncomfortable and at times, painful.⏎Importance

  4. #4 peak

    extract large amounts of caffeine using simple methods.⏎Now, caffeine has fairly good solubility in a variety of environments thanks to all those myriad polar and non-polar functional groups hanging off of it, which results in a wide range of possibilities for

  5. #5 peak

    activity in which they participate.⏎What are, in fact, motor skills? Well, this is a generic term that refers to fundamental movement skills, object control, stability factors and, last but not least, reaction speed. Countless sporting activities and

  6. #6 peak

    to maintain their current weight is to eat a well balanced diet composed of a variety of foods. To lose weight, your body needs a variety of nutrients. This can be accomplished by eating a

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 am going to use a thread to make my own image. What does this mean in technical terms? A A. An array of strings B. A set of boundaries in a force field C. A set of boundaries 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

    Western history isn’t littered with cravens and rapscallions.⏎As John Gallagher notes in Romania After Ceausescu, leaders such as Milosevic and Ceausescu used nationalism in ways that “denied individualism

  2. #2 peak

    ⏎Overview, Multiple objectives, Multiple stakeholders, Decision-making, Flood management⏎Civil and Environmental Engineering⏎Akter, Taslima and Simonovic, Slobodan P., "A General Overview of Multi-objective Multiple-participant Decision Making

  3. #3 peak

    . Antibiotics may also be given, even if the person does not have symptoms.⏎- Wash your hands often. Wash your hands<|endoftext|>Online, Maria Droujkova has founded the Natural Math site, the Natural Math Google group, the Math

  4. #4 peak

    say that many people have some subclinical ADHD-like symptoms but a diagnosis is only made once the symptoms become so prominent that they interfere with our everyday lives."⏎Predrag Petrovic, Consultant, Associate Professor, Psychiatry, Department of Clinical Neuroscience, Kar

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. L8:2217 excites
  2. L8:1397 excites
  3. L1:2620 excites
  4. L7:221 weakly suppresses
  5. L5:1181 weakly suppresses
  6. L8:378 weakly excites
  7. L7:1290 weakly excites
  8. L0:2033 weakly excites
  9. L6:2566 weakly excites
  10. L3:1633 weakly suppresses
  11. L5:1992 weakly suppresses
  12. L8:2966 weakly excites
  13. L6:2429 weakly suppresses
  14. L6:2898 weakly excites
  15. L5:1838 weakly suppresses
  16. L7:412 weakly excites
  17. L3:38 weakly excites
  18. L0:1480 weakly suppresses
  19. L5:1312 weakly suppresses
  20. L7:1922 weakly excites
  21. L2:458 weakly excites
  22. L8:2840 weakly suppresses
  23. L8:2680 weakly suppresses
  24. L3:2567 weakly suppresses

Feeds later layers

  1. L18:1361 suppresses
  2. L10:855 suppresses
  3. L17:1074 excites
  4. L17:915 suppresses
  5. L10:1486 suppresses
  6. L23:1450 excites
  7. L22:1754 suppresses
  8. L10:1411 excites
  9. L10:670 excites
  10. L20:2641 suppresses
  11. L23:2530 weakly suppresses
  12. L27:2448 weakly suppresses
  13. L10:870 weakly excites
  14. L21:1216 weakly excites
  15. L22:2550 weakly suppresses
  16. L21:1109 weakly suppresses
  17. L20:2063 weakly suppresses
  18. L18:3020 weakly excites
  19. L27:2478 weakly suppresses
  20. L22:2199 weakly suppresses
  21. L27:2879 weakly suppresses
  22. L10:1283 weakly suppresses
  23. L10:2894 weakly suppresses
  24. L26:2275 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 -2.83 and -2.721 values between -2.72 and -2.601 values between -2.60 and -2.481 values between -1.66 and -1.542 values between -1.31 and -1.192 values between -1.19 and -1.073 values between -1.07 and -0.963 values between -0.96 and -0.843 values between -0.84 and -0.723 values between -0.72 and -0.615 values between -0.61 and -0.4918 values between -0.49 and -0.3761 values between -0.37 and -0.25290 values between -0.25 and -0.14904 values between -0.14 and -0.021406 values between -0.02 and 0.10775 values between 0.10 and 0.22295 values between 0.22 and 0.3381 values between 0.33 and 0.4539 values between 0.45 and 0.576 values between 0.57 and 0.682 values between 0.68 and 0.801 values between 0.80 and 0.921 values between 1.04 and 1.152 values between 1.39 and 1.512 values between 1.74 and 1.86offnegativepositiveoften

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. battery
  2. 电池
  3. 万户
  4. battery
  5. 灯
  6. lights
  7. Battery
  8. fire
  9. Battery
  10. binary

Suppresses

  1. pj
  2. PCS
  3. tie
  4. 纶
  5. VEL
  6. &P
  7. qus
  8. PTH
  9. doch
  10. otch