Neuron Explorer

Qwen3-0.6B

mixed: names, 'for' and 'that' after nouns

A neuron in layer 21 of Qwen3-0.6B. Its strongest snippets are mixed: they peak on different tokens. It is featured, so it has steering outputs.

What this neuron does

This neuron is hard to pin down. It sometimes fires inside organisation names like 'International Union for Conservation' or inside a split-up name, and it goes quiet or negative on words like 'that' after a noun. Turning it up or down gave no clear theme.

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

    61 and 1964 by biologists in Russia. Hence, it is also called as the Russian Red Data Book.⏎Currently, the International Union for Conservation of Nature maintains the Red Data Book. IUCN is the world’s

  2. #2 peak

    at Macquarie University, Sydney, Australia.⏎Browse the Lab Book to learn about the history

  3. #3 peak

    global action under the Montreal Protocol to mend the hole in the ozone layer is starting to work. In the first accurate assessment of the impact of the treaty, Macquarie University climate dynamicist Murry Salby and his colleague Lilia Desch

  4. #4 peak

    ⏎Knowing which genetic variants are responsible for differences in gene activity among individuals can give powerful clues for diagnosis, prognosis and intervention of different diseases. Senior author Emmanouil Dermitzakis, Louis Jeantet Professor at UNIGE,

  5. #5 peak

    ellied.” In moments a crowd of doctors and attendants surrounded the two<|endoftext|>what is class in java⏎In Java, a class is a fundamental blueprint that defines the objects included in it.As such, it is comprised of methods, variables,

  6. #6 peak

    fruit bat (Styloctenium mindorensis) is found only on the Philippine island of Mindoro. Courtesy of ASU's International Institute for Species Exploration.⏎To date some 1.8 million species have been described,

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: it seems that a lot of people will do the same thing that does not seem to work. It will take the least amount of that which will do the best. It will take a non-mitestant, non-mitestant

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

    3×2×1 is written as 8! and we call it "eight factorial." You probably have a button on your calculator labelled with [x!] that will perform these calculations for you.⏎When you shuffle a pack of cards, you

  2. #2 peak

    You merely print it out, copy the graph paper more than on to your computer, or utilize it on a electronic graph plotter. There are even programs available that permit you to print it correct out of your computer with out a printer.⏎Numer

  3. #3 peak

    the first occurrence of value in the block of memory pointed by ptr.⏎If the value is not found, the function returns a null pointer.⏎In C, this function is only declared as:⏎void * memchr ( const void *, int, size

  4. #4 peak

    building a word frequency list and using your text to visualize your data, for the most part, in an easy and simple manner.⏎Antconc: Is freeware that is cross platform for Mac, Windows, and Linux. It is a concord

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. L19:2627 excites
  2. L15:2474 suppresses
  3. L20:1093 suppresses
  4. L20:2592 weakly suppresses
  5. L9:18 weakly suppresses text boundary after a document break
  6. L4:1208 weakly suppresses
  7. L20:96 weakly suppresses
  8. L17:973 weakly suppresses
  9. L20:1236 weakly excites
  10. L18:3061 weakly suppresses
  11. L18:1811 weakly excites 'and' linking two descriptive claims
  12. L12:1689 weakly excites
  13. L17:2489 weakly excites
  14. L20:1445 weakly excites
  15. L17:2779 weakly suppresses
  16. L15:1860 weakly excites
  17. L16:2512 weakly suppresses
  18. L13:135 weakly suppresses document boundary end-of-text token
  19. L19:2821 weakly suppresses
  20. L1:1872 weakly suppresses
  21. L18:1427 weakly excites
  22. L20:2778 weakly suppresses
  23. L19:1647 weakly excites
  24. L18:1151 weakly suppresses

Feeds later layers

  1. L22:335 suppresses
  2. L25:347 weakly suppresses
  3. L27:461 weakly suppresses
  4. L24:290 weakly suppresses 'to' in set phrases like 'to come'
  5. L23:638 weakly excites
  6. L24:156 weakly excites
  7. L23:2302 weakly suppresses
  8. L23:2079 weakly suppresses
  9. L22:1402 weakly suppresses
  10. L26:783 weakly suppresses unclear: word-start fragments like 'pl' and 'bl'
  11. L26:764 weakly suppresses
  12. L26:467 weakly suppresses
  13. L22:1171 weakly suppresses
  14. L22:181 weakly excites
  15. L25:586 weakly suppresses
  16. L26:246 weakly suppresses
  17. L25:812 weakly suppresses
  18. L23:1673 weakly excites
  19. L22:479 weakly suppresses
  20. L27:3015 weakly excites 'F' starting a word in a split
  21. L22:2958 weakly excites
  22. L22:402 weakly excites
  23. L25:2040 weakly suppresses
  24. L26:669 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 -18.30 and -17.741 values between -15.52 and -14.971 values between -13.30 and -12.741 values between -12.74 and -12.191 values between -11.08 and -10.522 values between -8.86 and -8.301 values between -8.30 and -7.751 values between -7.75 and -7.192 values between -7.19 and -6.642 values between -6.64 and -6.084 values between -6.08 and -5.531 values between -5.53 and -4.973 values between -4.97 and -4.425 values between -4.42 and -3.862 values between -3.86 and -3.306 values between -3.30 and -2.7511 values between -2.75 and -2.1910 values between -2.19 and -1.6421 values between -1.64 and -1.08162 values between -1.08 and -0.532138 values between -0.53 and 0.031449 values between 0.03 and 0.5863 values between 0.58 and 1.146 values between 1.14 and 1.697 values between 1.69 and 2.255 values between 2.25 and 2.802 values between 3.36 and 3.91offnegativepositiveoften

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. peq
  2. loo
  3. će
  4. 鼾
  5. representation
  6. れた
  7. .sep
  8. representing
  9. incare
  10. bi

Suppresses

  1. returns
  2. computes
  3. ives
  4. >Returns
  5. returns
  6. Returns
  7. приним
  8. reads
  9. utton
  10. 体会