Neuron Explorer

Qwen3-0.6B

unclear: mid-phrase words in prose

A neuron in layer 15 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 responds to a range of ordinary words in the middle of sentences, such as 'used', 'teach', 'trust' and 'country', with no clear shared meaning. The strongest cases are mostly plain prose rather than one kind of text. Turning it up changed the wording of the model's replies but not in any obvious direction.

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

    Springs, Ark.⏎Uses of radon⏎Radon is sometimes used<|endoftext|>A Stepper Motor is a special form of AC Motor. The input of

  2. #2 peak

    are used in medicine to teach clinical skills, diagnostic reasoning and planning treatment of individual patients. The way of making

  3. #3 peak

    pires of Ghana, Mali, and Songhai, and slave exports had supplemented the export of gold.⏎Paraguay and Panama were the countries with the highest level of trust in the

  4. #4 peak

    electoral college in the United States, it assigns electoral-college votes to each country (including the United States), based on the size of its population. So,

  5. #5 peak

    the good work and remember any<|endoftext|>07 Nov Book Review: Jim Crow Terminals⏎Book Review: Jim Crow Terminals:⏎The Desegregation of American Airports⏎Jacob Anbinder, Harvard University⏎Anke Ortlepp.

  6. #6 peak

    in 'Ylam at the eastern bank of today's Shatt-il-Arab.⏎These remarks indicate two ingredients in the formation of early Mesopotamian⏎Christianity, that have merged gradually into the general Christian population.⏎But we can state

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: I have been thinking about the other one: the one: the one: the one: the one: the one: the one: the one: the: the one: the: the one: the one: the one: 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

    and western Asia, but it was introduced and naturalized to Australia, New Zealand, and North America over time.⏎Want to save this post for later? Click Here to Pin It On Pinterest!⏎Are you interested in learning more about this ancient medicinal

  2. #2 peak

    your next visit.<|endoftext|>Sometimes it makes sense for a function to return multiple things. In such cases, it's common to just return them in a tuple: "return (count, new_obj)".⏎Sometimes you might want to return objects of

  3. #3 peak

    This will disconnect your terminal from the server you are currently connected to.⏎DWAINE is the operating system of Research Mainframes. Scientists use it to run their research programs. Research Directors program guardbuddies on DWAINE.⏎DWA

  4. #4 peak

    .⏎That is a lot of hardware and it’s also a lot of history. NASA is worried that without some guidance and agreement irreplaceable relics of the Space Age, such as Neil Armstrong’s first footprint on the Moon could be lost and sites

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. L14:2250 suppresses
  2. L14:2574 excites
  3. L14:1892 excites
  4. L14:110 excites
  5. L12:1373 suppresses
  6. L11:1588 excites
  7. L13:2917 excites
  8. L13:253 excites
  9. L13:1156 weakly suppresses
  10. L11:1122 weakly suppresses
  11. L13:397 weakly excites
  12. L11:2136 weakly excites
  13. L12:1712 weakly excites
  14. L13:219 weakly excites
  15. L14:1073 weakly suppresses
  16. L14:1638 weakly excites
  17. L14:2655 weakly excites
  18. L14:1434 weakly suppresses
  19. L13:1119 weakly excites
  20. L14:565 weakly excites
  21. L8:591 weakly excites
  22. L10:72 weakly suppresses
  23. L5:216 weakly excites
  24. L12:2946 weakly suppresses

Feeds later layers

  1. L16:3013 strongly suppresses
  2. L16:2393 strongly suppresses
  3. L16:1695 suppresses
  4. L18:2307 suppresses
  5. L16:1146 suppresses
  6. L18:761 suppresses
  7. L17:1323 suppresses
  8. L20:1595 suppresses
  9. L16:1578 suppresses
  10. L17:2271 suppresses
  11. L20:604 suppresses
  12. L21:349 suppresses
  13. L18:2205 suppresses
  14. L20:58 excites
  15. L19:2261 suppresses
  16. L16:2450 suppresses
  17. L18:1083 weakly suppresses
  18. L19:601 weakly suppresses
  19. L17:748 weakly suppresses
  20. L16:1474 weakly suppresses
  21. L20:484 weakly suppresses
  22. L17:614 weakly suppresses
  23. L17:1419 weakly suppresses
  24. L16:1108 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 -1.79 and -1.664 values between -1.53 and -1.402 values between -1.40 and -1.272 values between -1.27 and -1.148 values between -1.14 and -1.0119 values between -1.01 and -0.8816 values between -0.88 and -0.7535 values between -0.75 and -0.6276 values between -0.62 and -0.49137 values between -0.49 and -0.36227 values between -0.36 and -0.23503 values between -0.23 and -0.101263 values between -0.10 and 0.03565 values between 0.03 and 0.16351 values between 0.16 and 0.29195 values between 0.29 and 0.42141 values between 0.42 and 0.5592 values between 0.55 and 0.6877 values between 0.68 and 0.8154 values between 0.81 and 0.9437 values between 0.94 and 1.0733 values between 1.07 and 1.2021 values between 1.20 and 1.3311 values between 1.33 and 1.469 values between 1.46 and 1.597 values between 1.59 and 1.723 values between 1.72 and 1.853 values between 1.85 and 1.982 values between 1.98 and 2.113 values between 2.11 and 2.242 values between 2.24 and 2.372 values between 2.37 and 2.504 values between 2.50 and 2.631 values between 2.89 and 3.021 values between 3.02 and 3.151 values between 3.28 and 3.41offnegativepositiveoften

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. 把她
  2. 她说
  3. 告诉她
  4. 给她
  5. Gratuit
  6. ServiceImpl
  7. 主义者
  8. Giul
  9. 大发
  10. 零售商

Suppresses

  1. CD
  2. nd
  3. 0
  4. ometric
  5. cd
  6. los
  7. mp
  8. ert
  9. C
  10. ild