Neuron Explorer

Qwen3-0.6B

end-of-text token after finished articles

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

This neuron mostly fires at the point where one web page or article ends and a new unrelated one begins. It also fires on a few ordinary word breaks in news-style writing. Its negative side is less clear, showing up on words like 'their' or 'the' in the middle of a sentence.

Written by Claude from this neuron's activations and steering outputs. Confidence: low. 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

    pain) and seizures in some instances<|endoftext|>Gauss Sum Walks⏎This Demonstration shows two types of pseudorandom

  2. #2 peak

    source of clean water, soap so they can stay clean, and a proper toilet to avoid contaminating water supplies. In many of Oxfam’s ongoing programs, our partners work on promoting good hygiene and sanitation to help people stay healthy even when

  3. #3 peak

    turtles from the wild and sell them.) Because of her size, she was rescued and sent to a Missouri farm, which is where the Newport Aquarium found her.<|endoftext|>The hustle and bustle of the back-to-school season is starting to wind down

  4. #4 peak

    York’s Metropolitan Museum of Art in 1983. The Designer died in Paris on June 1, 2008 from brain cancer.<|endoftext|>How do you get

  5. #5 peak

    the U.S. Congress to agree on a carbon tax or cap, his idea has serious merit and deserves a broad discussion, especially as Congress reassess its budget priorities. This includes billions of dollars of tax breaks and incentives for oil and other fossil

  6. #6 peak

    Unfortunately he died at a very young age, in 1938, which made it impossible for him to leave and extensive body of work behind. .<|endoftext|>|Siege of Jerusalem|⏎|Part of the First Jewish-Roman War|⏎

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 first thing in the question is the time, and the first thing is the time when it is, and the first thing is the time, and so, the first thing is the time, and the first is the time, and 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

    better!)⏎Newspapers and magazines⏎Writing utensils (markers, pencils)⏎Glue or tape⏎- Gather and look at newspapers for layout inspiration. Consider their design, the placement of

  2. #2 peak

    and they had two children. Just a few years later, the young mother traveled to Rome to study under Maria Montessori. Impressed by Montessori’s revolutionary methods, Fisher worked to bring the Montessori Method to the United States

  3. #3 peak

    "?⏎Yes, curling has been compared to chess by fans and detractors alike. Fans emphasize the strategy involved. Detractors claim that both sports are similarly long and tedious for

  4. #4 peak

    such as a mobile or through your hotel.⏎- If you receive an email which appears to be from a family member or friend claiming to be in trouble look at the way the email is written, ask yourself whether the email sounds like it was written

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. L12:3069 suppresses
  2. L10:1884 excites
  3. L11:1092 suppresses
  4. L12:2280 excites pipe before a table field label
  5. L9:941 excites
  6. L12:1794 excites
  7. L12:1913 excites
  8. L11:1532 suppresses
  9. L11:2765 excites
  10. L10:2293 suppresses
  11. L12:2414 suppresses
  12. L12:1824 weakly excites
  13. L6:623 weakly suppresses
  14. L12:2201 weakly suppresses
  15. L12:1518 weakly suppresses
  16. L12:1838 weakly excites
  17. L12:1121 weakly excites
  18. L10:331 weakly excites
  19. L12:1828 weakly suppresses
  20. L12:3032 weakly suppresses
  21. L1:171 weakly excites "but" after a negated clause
  22. L9:2721 weakly suppresses
  23. L10:2522 weakly excites
  24. L12:1432 weakly excites

Feeds later layers

  1. L14:252 suppresses
  2. L15:1369 suppresses
  3. L16:2823 suppresses
  4. L16:2988 suppresses
  5. L14:1664 suppresses
  6. L16:733 suppresses
  7. L15:209 suppresses
  8. L15:2265 suppresses
  9. L14:2813 suppresses
  10. L17:2565 suppresses
  11. L15:266 excites
  12. L16:3061 suppresses
  13. L17:2554 suppresses
  14. L15:1998 suppresses
  15. L18:2261 suppresses
  16. L16:2010 suppresses
  17. L14:2290 suppresses
  18. L18:2548 suppresses
  19. L17:2374 weakly suppresses
  20. L14:2356 weakly suppresses
  21. L18:2147 weakly suppresses
  22. L18:1358 weakly suppresses
  23. L16:3019 weakly suppresses
  24. L15:709 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 -2.53 and -2.411 values between -2.41 and -2.291 values between -1.94 and -1.823 values between -1.46 and -1.343 values between -1.34 and -1.221 values between -1.22 and -1.109 values between -1.10 and -0.985 values between -0.98 and -0.8613 values between -0.86 and -0.7417 values between -0.74 and -0.6226 values between -0.62 and -0.5045 values between -0.50 and -0.3890 values between -0.38 and -0.26205 values between -0.26 and -0.14788 values between -0.14 and -0.021871 values between -0.02 and 0.10507 values between 0.10 and 0.22190 values between 0.22 and 0.3463 values between 0.34 and 0.4627 values between 0.46 and 0.5814 values between 0.58 and 0.7012 values between 0.70 and 0.823 values between 0.82 and 0.945 values between 0.94 and 1.054 values between 1.05 and 1.172 values between 1.17 and 1.292 values between 2.13 and 2.25offnegativepositiveoften

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. itionally
  2. usk
  3. 得住
  4. alyzer
  5. ONGL
  6. enko
  7. 울
  8. otland
  9. createContext
  10. yn

Suppresses

  1. features
  2. feature
  3. features
  4. Latitude
  5. _feature
  6. eam
  7. Feature
  8. feature
  9. eras
  10. 封面