Neuron Explorer

Qwen3-0.6B

unclear: mid-sentence word choices, mixed

A neuron in layer 16 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 fires on a grab bag of words in the middle of sentences, such as a number in a list, the word 'five', or 'true' in a comparison. I could not find one clear theme. Turning it up made the model ramble about unrelated topics.

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

    do not contain zeros: 1, 2, 3, 4, 5, 6, 7, 8, 9, 13, 14, 15, 16,

  2. #2 peak

    plants, and a virus blows in across the ocean from a remote island, your lawn has a better chance of surviving and renewing itself than if you have only five kinds of plants. In the same way, having more species of wild birds will

  3. #3 peak

    no", "yes"]⏎Note that the arguments to⏎bool are in the opposite order of the⏎if then else-syntax; First the false value, then the true value, and finally the boolean. This argument order is consistent with the data type

  4. #4 peak

    than two numbers are being multiplied, just count the number of minus signs: any even number of minus signs collapses to a plus sign; any odd number of minus signs collapses to a minus sign.⏎Input your answer as:

  5. #5 peak

    ātauranga strategy for Ngāti Kahungunu are as follows:⏎Strengthen, support and guide Kahungunu ākonga, whā

  6. #6 peak

    classified by sex in the form of a histogram, it is called age-sex pyramid. The base of the pyramid shows very low age starting from zero, it increases as we move upwards. The top of the pyramid shows the maximum age above 8

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 set of all sets that are a set of some set of some set of some set of some set of some set of ... the set of some set of some set of ... the set of some set of some set of some set of

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

    a branch of mathematics studying zeros of polynomial equations. Modern algebraic geometry is based on more abstract techniques of abstract algebra, especially commutative algebra, with the language and the problems of geometry.⏎The main objects of study

  2. #2 peak

    work for arbitrary toposes can be deduced by arguing from universal elements. Some simple examples follow.⏎Consider first the construction of internal conjunction .⏎is equivalent to the nerve of .⏎As a warm-up:⏎The colimit of⏎is isomorphic to

  3. #3 peak

    “Complex Geometry“:⏎In mathematics, complex geometry is the study of complex manifolds and functions of many complex variables. Application of transcendental methods to algebraic

  4. #4 peak

    classically Diophantine geometry), a subfield of algebraic number theory.⏎- The study of the real points of an algebraic variety is

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. L9:65 suppresses
  2. L3:25 suppresses
  3. L10:276 excites
  4. L3:461 excites
  5. L7:334 excites
  6. L9:41 excites
  7. L6:1525 excites
  8. L5:221 suppresses
  9. L3:924 suppresses
  10. L9:693 suppresses
  11. L10:2217 suppresses
  12. L5:78 excites
  13. L9:96 excites
  14. L9:1781 excites
  15. L0:281 excites
  16. L10:544 suppresses
  17. L1:137 suppresses
  18. L3:105 excites
  19. L3:448 suppresses
  20. L3:63 excites
  21. L6:1459 excites
  22. L8:77 excites
  23. L9:487 excites
  24. L3:408 weakly excites

Feeds later layers

  1. L21:246 suppresses
  2. L20:153 suppresses
  3. L24:263 suppresses
  4. L26:330 suppresses
  5. L26:173 suppresses end-of-document boundary token
  6. L21:987 suppresses
  7. L24:657 suppresses
  8. L21:345 suppresses
  9. L25:299 suppresses
  10. L24:246 suppresses
  11. L21:1762 suppresses
  12. L21:684 suppresses
  13. L24:287 suppresses
  14. L22:1196 suppresses
  15. L24:310 suppresses
  16. L23:1330 suppresses
  17. L19:1027 suppresses
  18. L20:461 suppresses
  19. L20:157 suppresses
  20. L25:2827 suppresses
  21. L21:947 suppresses
  22. L21:630 suppresses
  23. L20:870 suppresses
  24. L22:1080 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 -9.42 and -9.142 values between -1.43 and -1.154 values between -1.15 and -0.8618 values between -0.86 and -0.57179 values between -0.57 and -0.291058 values between -0.29 and -0.001669 values between -0.00 and 0.28596 values between 0.28 and 0.57239 values between 0.57 and 0.8590 values between 0.85 and 1.1432 values between 1.14 and 1.4215 values between 1.42 and 1.715 values between 1.71 and 1.99offnegativepositiveoften

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. 始建
  2. 通行证
  3. 首家
  4. USER
  5. lemetry
  6. wayne
  7. webkit
  8. civilian
  9. MainThread
  10. 新一轮

Suppresses

  1. leurs
  2. ators
  3. -unit
  4. ören
  5. uations
  6. jections
  7. otomy
  8. erator
  9. PermissionsResult
  10. capit