Neuron Explorer

Qwen3-0.6B

'of' or 'to' linking two things

A neuron in layer 13 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

Fires on the small word right after a noun that ties it to another thing, like 'a division of', 'a larger version of', 'a special case of' or 'related to'. Turning it up made the model produce loose, repetitive text, which does not clearly show what it is for.

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

    to get integer outputs in all cases⏎Determine value using Look-up table method⏎Create Volcano (or Atoll) martix. It is an extension of Bullseye matrix problem.⏎Who has power to do everything in this world?⏎

  2. #2 peak

    downside: Your new pc comes preloaded with applications.⏎Phrase Origin late Center English: the verb from Old French practiser or medieval Latin practizare, alteration of practicare ‘perform, perform’, from practica ‘apply’, from Greek prakt

  3. #3 peak

    ure; to accustom by practice.⏎from The Century Dictionary and Cyclopedia⏎- To work; practise; inure; exercise.⏎- A Middle English form of our.⏎- n. Operation; use; practice.⏎- n. Fortune;

  4. #4 peak

    Highway Administration, a division of the U.S. Department of Transportation, has released the newest numbers for miles driven

  5. #5 peak

    of the Nieuport 12 with an 80hp Le Rhône engine. The Nieuport 12 itself was a larger version of the Nieuport 10, with a more powerful 110

  6. #6 peak

    squirrels, tamarins are commonly called kings of the jungle because manes around their neck make them look like lions. The lion tamarins are related to pygmy marmosets. They are 12 inches tall not

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: the problem of a single problem, which is the first, in the first place, as one of the first problems in the first place, and the second, and the third, in the first place, as a consequence of the first,

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

    cells.⏎The distance between two cells is the minimal number of up, down, left or right steps through the grid to get from one to the other. It doesn't matter whether the steps go through blank cells or

  2. #2 peak

    Add your thoughts below. You can sign in with Google, Twitter, Facebook, Yahoo! to get

  3. #3 peak

    they explore side-branches, they make mistakes. The “right answer” – the proof, the demonstration – simply sets a mile-marker between one phase of discovery and the next.⏎In our Math Circles, young people invent and discover their shared

  4. #4 peak

    about the art of restoration. It is usually referred to as the non-institutional knowledge passed from master craftsman to apprentice, father to son, or artist to pupil. This is the main difference between restoration and conservation. Both restoration and conservation

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. L6:2605 suppresses
  2. L11:2292 excites 'than' in comparisons
  3. L12:838 excites
  4. L9:1144 excites
  5. L6:1477 suppresses
  6. L12:2959 suppresses
  7. L12:2298 excites
  8. L8:1319 suppresses
  9. L12:2548 excites
  10. L10:3066 suppresses
  11. L11:139 excites
  12. L7:2797 excites
  13. L6:1084 suppresses
  14. L6:1828 suppresses
  15. L10:2213 weakly suppresses 'to' or 'into' showing movement between two things
  16. L9:2566 weakly excites
  17. L11:2478 weakly excites
  18. L12:1987 weakly excites
  19. L11:2649 weakly excites
  20. L8:1669 weakly excites
  21. L12:2346 weakly excites
  22. L8:1623 weakly excites
  23. L11:2662 weakly suppresses
  24. L9:2885 weakly excites

Feeds later layers

  1. L14:2386 weakly suppresses
  2. L18:1738 weakly suppresses
  3. L16:1535 weakly excites
  4. L17:2920 weakly excites
  5. L17:2769 weakly suppresses
  6. L18:2373 weakly suppresses
  7. L23:1239 weakly suppresses
  8. L19:103 weakly excites
  9. L18:1420 weakly suppresses
  10. L14:2535 weakly suppresses
  11. L16:765 weakly excites
  12. L18:860 weakly excites
  13. L17:1467 weakly suppresses
  14. L23:1889 weakly suppresses the '3' in '3D' and '2D'
  15. L16:980 weakly excites
  16. L17:2516 weakly excites
  17. L14:2374 weakly excites
  18. L14:2422 weakly excites
  19. L14:363 weakly excites
  20. L15:1126 weakly suppresses
  21. L17:439 weakly excites
  22. L22:418 weakly suppresses
  23. L14:229 weakly excites
  24. L14:1654 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 -1.81 and -1.651 values between -1.48 and -1.314 values between -1.31 and -1.157 values between -1.15 and -0.9810 values between -0.98 and -0.8121 values between -0.81 and -0.6447 values between -0.64 and -0.4891 values between -0.48 and -0.31271 values between -0.31 and -0.141880 values between -0.14 and 0.021250 values between 0.02 and 0.19239 values between 0.19 and 0.3643 values between 0.36 and 0.5312 values between 0.53 and 0.696 values between 0.69 and 0.868 values between 0.86 and 1.033 values between 1.03 and 1.193 values between 1.19 and 1.361 values between 1.36 and 1.534 values between 1.53 and 1.702 values between 2.03 and 2.202 values between 2.36 and 2.531 values between 4.04 and 4.201 values between 4.70 and 4.87offnegativepositiveoften

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. aan
  2. magic
  3. ồn
  4. oted
  5. permanent
  6. -exec
  7. otent
  8. odynam
  9. uada
  10. Insn

Suppresses

  1. lessness
  2. nonetheless
  3. ompiler
  4. ллер
  5. ǧ
  6. jährige
  7. jähr
  8. &q
  9. 公报
  10. '&