Neuron Explorer

Qwen3-0.6B

unclear: end of document token

A neuron in layer 19 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

The neuron's strongest reactions are around the marker that separates one web page from the next in the training text, which the model sees as a hard stop. Its large reactions are negative, and it does nothing clear when turned up.

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

    . It will contain an appropriate error⏎message for the user.⏎Upon success, this is the file handle returned by⏎``urllib2.urlopen``. Per urllib2, it has an ``info()`` method⏎containing things like headers

  2. #2 peak

    Increases Cardiovascular Risks⏎Led by James L. Vacek, M.D., a team of researchers at the University of Kansas Medical Center undertook a<|endoftext|>The Tenderlett lancet was made for single use in a lancing device.

  3. #3 peak

    the China site (in Chinese or English) for best site performance. Other MathWorks country sites are not optimized for visits from your location.⏎Contact your local office<|endoftext|>It’s PrAACtically Martin Luther King Jr Day⏎Looking for activities and

  4. #4 peak

    , or whether it’s heavy metals, or whether it’s nitrogen and phosphorus – will make it to the sea. And that means that probably water quality will<|endoftext|>Harlem riot of 1935⏎The Harlem riot of 1

  5. #5 peak

    it encountered the overwhelming gravitational force of the much larger mass. This force wrenched luminous stars from the small galaxy. Over billions of years of evolution, the<|endoftext|>Duplicating Data Elements⏎Using keywords and synonyms is one way to ensure that

  6. #6 peak

    the China site (in Chinese or English) for best site performance. Other MathWorks country sites are not optimized for visits from your location.⏎Contact your local office<|endoftext|>The point in the Moon's orbit where it lies diametrically⏎opposite

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 time I met with the first #1 #2 #3 #4 #1 #4 #1 #1 #1 #1 #1 #1 #1

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

    'S', then click at the end of the word and type<|endoftext|>Mallorca’s popular tourist railway, the ‘Orange Express,’ has entered its

  2. #2 peak

    and Ragusa.<|endoftext|>Yellow fever is a disease carried by female mosquitoes. The species of mosquito that carry

  3. #3 peak

    running. Slow to<|endoftext|>South-eastern Australia is one of the most bushfire-prone areas in the

  4. #4 peak

    comment.<|endoftext|>100 Chemistry Questions and Answers 8211; Basic Chemistrynbsp

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. L9:18 suppresses text boundary after a document break
  2. L9:114 suppresses
  3. L10:32 excites
  4. L17:644 suppresses
  5. L8:341 suppresses
  6. L18:109 suppresses
  7. L3:163 excites
  8. L3:5 excites
  9. L4:382 suppresses
  10. L17:1334 excites
  11. L18:555 excites end-of-text marker between documents
  12. L9:948 excites
  13. L9:49 suppresses
  14. L7:467 excites
  15. L1:1035 weakly excites
  16. L2:1092 weakly excites
  17. L5:2 weakly suppresses
  18. L7:453 weakly excites
  19. L14:67 weakly suppresses
  20. L15:570 weakly suppresses
  21. L10:1772 weakly excites
  22. L2:322 weakly excites
  23. L9:1130 weakly suppresses
  24. L5:1007 weakly excites

Feeds later layers

  1. L21:282 suppresses
  2. L20:96 excites
  3. L25:436 suppresses
  4. L22:221 suppresses period ending a contact or help sentence
  5. L26:1283 suppresses
  6. L21:955 suppresses
  7. L22:191 suppresses
  8. L27:1300 suppresses
  9. L27:256 suppresses
  10. L26:120 suppresses
  11. L26:601 suppresses
  12. L26:228 suppresses
  13. L22:230 weakly suppresses
  14. L21:652 weakly suppresses
  15. L22:2447 weakly excites
  16. L23:813 weakly excites
  17. L20:255 weakly suppresses
  18. L20:116 weakly suppresses
  19. L27:118 weakly suppresses
  20. L22:640 weakly suppresses
  21. L24:161 weakly suppresses
  22. L22:2374 weakly suppresses
  23. L21:633 weakly suppresses
  24. L26:59 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 -30.05 and -29.241 values between -19.58 and -18.771 values between -14.74 and -13.941 values between -9.11 and -8.302 values between -7.50 and -6.693 values between -6.69 and -5.882 values between -5.88 and -5.083 values between -4.27 and -3.476 values between -3.47 and -2.667 values between -2.66 and -1.8619 values between -1.86 and -1.0560 values between -1.05 and -0.251995 values between -0.25 and 0.561745 values between 0.56 and 1.3762 values between 1.37 and 2.17offnegativepositiveoften

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. anced
  2. …”
  3. …)
  4. +]
  5. arks
  6. ographers
  7. \"]
  8. =").
  9. ”).
  10. :\"

Suppresses

  1. import
  2. ⏎ ⏎ ⏎
  3. ⏎⏎⏎
  4. https
  5. ⏎
  6. ⏎ ⏎
  7. ⏎ ⏎ ⏎ ⏎
  8. ⏎⏎⏎
  9. ⏎⏎⏎⏎