Neuron Explorer

Qwen3-0.6B

hair and hair anatomy words

A neuron in layer 23 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 'hair' and on nearby words in texts about hair, hair structure or plant parts. When turned up, the model started writing about roots and root systems.

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

    transparent, overlapping scales, it resembles the shingles on a roof. The cuticle's primary purpose is to safeguard the inner structures of the hair, such as the cortex and medulla, from damage caused by external factors like heat, chemicals,

  2. #2 peak

    for our landfill.⏎There are lots of products you can switch to:⏎- Instead of bottled water, change for tap water in a water bottle⏎- Swap your hair washing products for shampoo and conditioner bars⏎- Say goodbye to shower gel and hello

  3. #3 peak

    , WWI Map Is Up for Auction⏎- Thousands Of FBI Documents About Civil Rights Era Destroyed By Flooding⏎- Ancient Egyptian Woman with 70 Hair Extensions Discovered⏎- Europeans drawn from three ancient 'tribes'⏎- Conservatives

  4. #4 peak

    affected by electrical noise emitted by gadgets like smartphones, TVs, computers, even vacuum cleaners and hairdryers, AM signals are prone to distortion and crackling.⏎That crackle

  5. #5 peak

    and personal-care products (soaps, shampoos, hair sprays, and nail polishes).⏎Phthalates are used widely in poly

  6. #6 peak

    .⏎- What is the turkey's beard made of? Turkey beards are actually comprised of bristles or filaments that appear to be hair-like, modified feathers known as mesofiloplumes. The individual bristles emerge

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 can't find the map to the first two books in the book map and I need to find a way to move around within the book map. The first two books are important to me, so I need to understand the pattern of the

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

    the top to 5 lobes, the lobes longer than the tubes and becoming reflexed.⏎Fruits: Berrylike, globose, 5-6 mm across, shiny red or purplish-black, occasionally brown, yellow,

  2. #2 peak

    ile, style 1, entire or 3-branched⏎Fruit: capsule, loculicidal, thin walled, ± rounded, 6-angled, or winged, chambers 3⏎Seeds many,

  3. #3 peak

    Citrus reticulata) is a tree from the Rutaceae family producing citrus fruits called mandarins. The fruit is spherical and slightly flattened, with a diameter of 5 to 9 cm. Its flesh, sweet and frag

  4. #4 peak

    vate, mid rib conspicuous, apex with narrowly inflexed lobule. Stylopodium conic; styles very short, reflexed. Fruit oblong, terete, glabrous; ribs 5, acute or round-obt

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. L20:1268 suppresses
  2. L21:2902 suppresses
  3. L9:3018 weakly excites
  4. L17:1534 weakly excites
  5. L7:204 weakly excites
  6. L20:2644 weakly excites
  7. L20:963 weakly suppresses
  8. L22:1122 weakly suppresses
  9. L19:1324 weakly suppresses
  10. L8:1612 weakly excites
  11. L11:2202 weakly suppresses
  12. L5:1267 weakly suppresses
  13. L2:483 weakly suppresses
  14. L10:852 weakly excites
  15. L19:2860 weakly suppresses
  16. L5:216 weakly suppresses
  17. L6:1413 weakly excites
  18. L2:1321 weakly suppresses
  19. L4:2285 weakly excites
  20. L0:2609 weakly suppresses
  21. L22:2106 weakly excites
  22. L18:2783 weakly suppresses
  23. L18:1460 weakly suppresses
  24. L22:254 weakly suppresses

Feeds later layers

  1. L25:2202 weakly excites
  2. L25:2725 weakly suppresses
  3. L25:2037 weakly excites
  4. L25:84 weakly excites
  5. L26:1543 weakly excites
  6. L24:2635 weakly excites
  7. L27:2534 weakly excites
  8. L25:1610 weakly excites
  9. L24:1947 weakly excites
  10. L27:1668 weakly excites
  11. L24:712 weakly excites 'long' in 'long-term' and 'long-lasting'
  12. L24:3013 weakly excites
  13. L24:2229 weakly suppresses
  14. L24:1418 weakly excites
  15. L24:1892 weakly suppresses
  16. L25:2894 weakly excites
  17. L24:69 weakly excites
  18. L26:2844 weakly excites
  19. L25:1290 weakly suppresses
  20. L26:2000 weakly excites
  21. L24:615 weakly excites
  22. L25:2950 weakly suppresses 'in' in 'in progress' phrases
  23. L25:2036 weakly excites
  24. L24:2280 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 -30.38 and -28.751 values between -23.85 and -22.221 values between -22.22 and -20.591 values between -18.96 and -17.334 values between -12.44 and -10.814 values between -10.81 and -9.181 values between -9.18 and -7.554 values between -7.55 and -5.928 values between -5.92 and -4.2921 values between -4.29 and -2.6663 values between -2.66 and -1.033506 values between -1.03 and 0.60201 values between 0.60 and 2.2341 values between 2.23 and 3.8715 values between 3.87 and 5.5012 values between 5.50 and 7.136 values between 7.13 and 8.768 values between 8.76 and 10.394 values between 10.39 and 12.023 values between 12.02 and 13.651 values between 13.65 and 15.281 values between 18.54 and 20.171 values between 33.21 and 34.84offnegativepositiveoften

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. 根
  2. 茎
  3. 枝
  4. root
  5. -root
  6. Branch
  7. heads
  8. boom
  9. (axis
  10. 直

Suppresses

  1. .checkSelfPermission
  2. че
  3. Subaru
  4. atches
  5. atcher
  6. RI
  7. essel
  8. isto
  9. éric
  10. iore