Neuron Explorer

Qwen3-0.6B

first letter m or oct fragments

A neuron in layer 25 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 first piece of a longer word that has been split into pieces, most often the letters 'm' or 'oct' starting words like 'mussels', 'mackerel' or 'octinoxate'. Turning it up made the model output long runs of blank lines. The opposite side is about words like 'over' and 'design' with no shared pattern.

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

    hu”. Since all the devotees who performed pooja on this auspicious day will achieve Moksha (Eternal Bliss), this day is also called “Mokshothasava Day”. Since all thirty three crore Gods are relived

  2. #2 peak

    and refuges (coral reefs) and provide food for fish (krill, pteropods, shellfish larvae) and humans (oysters, mussels, clams). Recent research reveals just how threatened such species are by acid

  3. #3 peak

    and ultimately make their way into the oceans. This is how the beauty industry affects our ocean ecosystem threatening marine life. Chemicals, especially oxybenzone, and octinoxate, found in sunscreen contribute to coral bleaching.⏎The Emergence of

  4. #4 peak

    asi Tidhis in a year Suklapaksha Ekadasi is termed as “Vaikuntha Ekadasi”. It is also called as “Mukkoti Ekadasi”. Though Dhanurmasa occurs in Sour

  5. #5 peak

    launching its new laptops known as Chromebooks and apparently, these do not contain any installed software.<|endoftext|>Salmon and other fatty fish such as sardines and mackerel are the superstars of heart-healthy foods. That’s because they

  6. #6 peak

    he was still a babe in his cradle and he strangled the two serpents that were meant to kill him, as the morning sun dispersed the creeping mists.⏎If Zeus-Order had ruled the world, the right of the first

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: how to make a healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy healthy

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

    outhern India frequently accessorize their traditional garment with a spare⏎scrap of cloth, called an angavastram. The garment is designed to be worn over⏎the shoulder. This garment will remain on the shoulders for the time being.⏎

  2. #2 peak

    ⏎the results of the first stage. As of September 2008 the design teams⏎use 3D modeling software to generate CAD drawings from initial design⏎sketches. The designs are then prototyped using an in-house

  3. #3 peak

    vioglacial waters. Up until 1860 amber procuring methods were off⏎the beaches and with shallow diving. It was obvious that sea amber came⏎from strata beneath the sea and a larger supply could be obtained by⏎

  4. #4 peak

    made certain recommendations on the basis of which the Chief⏎Ministers in their Conference held in July, 1972, laid down the ‘National Guide lines’⏎that were to govern the ceiling<|endoftext|>Daddy Long Legs and The Inch

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. L22:377 strongly excites
  2. L20:268 strongly excites
  3. L20:1277 strongly excites
  4. L22:165 strongly suppresses
  5. L18:528 strongly excites
  6. L19:1090 strongly suppresses
  7. L19:2773 strongly suppresses
  8. L22:751 strongly suppresses
  9. L9:808 strongly excites
  10. L7:74 strongly excites unclear: scattered mid-sentence tokens
  11. L8:380 strongly excites
  12. L16:442 strongly excites
  13. L21:579 strongly suppresses
  14. L20:515 strongly excites period right before a line break
  15. L15:643 strongly suppresses
  16. L18:702 strongly suppresses
  17. L17:696 strongly excites
  18. L19:973 strongly excites
  19. L16:3055 strongly suppresses
  20. L24:1619 strongly excites
  21. L21:1564 strongly excites
  22. L24:244 suppresses
  23. L14:568 suppresses
  24. L24:1434 excites sentence end before a line break

Feeds later layers

  1. L26:2953 excites
  2. L27:1017 excites
  3. L26:2873 excites
  4. L27:1536 excites
  5. L26:230 excites
  6. L26:757 excites
  7. L27:2557 excites
  8. L27:1681 excites
  9. L27:598 excites
  10. L27:116 excites
  11. L26:740 weakly excites
  12. L27:207 weakly suppresses
  13. L27:97 weakly suppresses
  14. L27:487 weakly excites
  15. L26:993 weakly excites
  16. L26:460 weakly excites
  17. L27:1801 weakly excites
  18. L26:758 weakly excites
  19. L26:167 weakly excites
  20. L26:2346 weakly suppresses
  21. L26:449 weakly suppresses
  22. L26:787 weakly suppresses
  23. L26:749 weakly excites
  24. L26:1018 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 -17.53 and -16.591 values between -12.81 and -11.861 values between -8.08 and -7.141 values between -7.14 and -6.193 values between -5.25 and -4.304 values between -4.30 and -3.362 values between -3.36 and -2.4111 values between -2.41 and -1.47349 values between -1.47 and -0.523146 values between -0.52 and 0.42256 values between 0.42 and 1.3777 values between 1.37 and 2.3119 values between 2.31 and 3.2615 values between 3.26 and 4.206 values between 4.20 and 5.153 values between 5.15 and 6.094 values between 6.09 and 7.042 values between 7.04 and 7.981 values between 7.98 and 8.931 values between 9.87 and 10.821 values between 10.82 and 11.762 values between 16.49 and 17.431 values between 17.43 and 18.381 values between 19.32 and 20.27offnegativepositiveoften

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. ⏎
  2. ⏎
  3. ⏎⏎
  4. ⏎
  5. " ⏎
  6. ' ⏎
  7. - ⏎
  8. ⏎ ⏎
  9. "⏎
  10. '⏎

Suppresses

  1. ниц
  2. .firebaseapp
  3. .BorderColor
  4. ño
  5. ymce
  6. roid
  7. ⍣
  8. lander
  9. ophile
  10. rement