Neuron Explorer

Qwen3-0.6B

'of' after geographic nouns

A neuron in layer 12 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 lights up on the word 'of' when it follows a place or landmark, like 'shores of Hawaii', 'floor of the mouth', or 'west coast of Mexico'. It is quiet on most other text. Turning it up gave repetitive sentences that were not clearly about geography.

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

    of the massive convergence of marine litter known as the Great Pacific Garbage Patch. The easternmost concentration of trash is midway between the California coast and the eastern shores of Hawaii.⏎These

  2. #2 peak

    the oral cavity is produced by the submandibular glands, even though they are much smaller than the .⏎The are a pair of glands located beneath the floor of the mouth anterior to the . The secretion produced is mainly in nature, however it

  3. #3 peak

    churban helps us mourn its destruction. Tisha B’Av is the culmination, of course, of two national tragedies: the loss of the first Beis HaMikdash as well as the Second Beis HaMikdash.⏎To

  4. #4 peak

    the worst. Learn more about Ninja Polymers in the infographic below, and share individual story sections by clicking on the embed icon in the upper right-hand corner of

  5. #5 peak

    shall begin on the Gulph [sic] of Mexico, at the mouth of the river Sabine, in the sea, continuing north, along the western bank of that river, to the 32d degree of latitude; thence,

  6. #6 peak

    way stretches 4,000 miles north-to-south and 1,000 miles east-to-west. From the Arctic to the west coast of Mexico and the Rocky Mountains to

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: [X]A and X]A, the same A, are [X]A (A)X (X) [A]X (A)X (A)A. (A)X (A)A (A

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

    , fellow skeptic, and fellow blogger Jeff Medkeff discovered the asteroid in 2000. It was given the preliminary designation of 2000 WG11, and Jeff had the privilege of naming it, and the short

  2. #2 peak

    For further information and to get help call the Poison Help Line at 1-800-222-1222.⏎Synonyms and

  3. #3 peak

    010] Please see my latest post about setting an ad hoc network to share an internet connection.<|endoftext|>Encyclopedia of Southern Jewish Communities⏎History of Congregation B'nai Israel, Little Rock⏎B’nai Israel congregation was first organized

  4. #4 peak

    're saving.⏎For More Information...⏎Phone: 1-888-STAR-YES (1-888-782-7937)⏎For more information on activities ongoing

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. L11:1246 excites
  2. L10:1748 excites
  3. L11:2953 suppresses
  4. L6:1563 excites
  5. L6:2664 excites
  6. L8:1544 excites
  7. L10:2786 excites
  8. L8:526 suppresses
  9. L8:2956 excites
  10. L8:502 excites
  11. L8:660 suppresses
  12. L9:1428 excites
  13. L10:2617 suppresses
  14. L8:1755 suppresses
  15. L10:1395 excites
  16. L8:1003 suppresses "the" or "of" before a noun phrase
  17. L6:1244 suppresses
  18. L11:1611 excites
  19. L6:1412 suppresses
  20. L5:2114 weakly suppresses unclear: calculus notation, mixed topics
  21. L10:1706 weakly suppresses
  22. L9:998 weakly suppresses
  23. L9:2764 weakly suppresses
  24. L10:272 weakly suppresses

Feeds later layers

  1. L14:1659 suppresses
  2. L13:856 suppresses
  3. L13:210 suppresses
  4. L16:3012 suppresses
  5. L20:331 excites
  6. L18:725 suppresses
  7. L18:370 suppresses
  8. L20:1112 suppresses
  9. L14:539 suppresses
  10. L16:229 suppresses
  11. L14:126 excites
  12. L13:2049 suppresses
  13. L20:1218 suppresses
  14. L16:1321 suppresses
  15. L19:414 suppresses
  16. L14:1971 suppresses
  17. L15:1299 suppresses
  18. L17:1240 excites
  19. L19:761 suppresses
  20. L14:2275 suppresses
  21. L14:655 suppresses
  22. L16:985 weakly suppresses
  23. L15:1973 weakly excites
  24. L17:264 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.

2 values between -4.12 and -4.001 values between -3.39 and -3.272 values between -2.41 and -2.291 values between -1.55 and -1.433 values between -1.31 and -1.194 values between -1.19 and -1.077 values between -1.07 and -0.942 values between -0.82 and -0.705 values between -0.70 and -0.5810 values between -0.58 and -0.4525 values between -0.45 and -0.3382 values between -0.33 and -0.21335 values between -0.21 and -0.091118 values between -0.09 and 0.041150 values between 0.04 and 0.16662 values between 0.16 and 0.28302 values between 0.28 and 0.40136 values between 0.40 and 0.5246 values between 0.52 and 0.6515 values between 0.65 and 0.77offnegativepositiveoften

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. nock
  2. ัย
  3. forfeiture
  4. =?";⏎
  5. .blogspot
  6. uelle
  7. столь
  8. edException
  9. 煲
  10. such

Suppresses

  1. Mass
  2. anything
  3. Var
  4. entimes
  5. inos
  6. Mass
  7. ((*
  8. olate
  9. 奥
  10. ovo