Neuron Explorer

Qwen3-0.6B

unclear: mixed mid-sentence word choices

A neuron in layer 7 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 reacts to a grab bag of words in ordinary English sentences, such as 'reasonable', 'clear' and 'skills', with no single theme we could find. It tends to go quiet on code and list-like text. Turning it up did not produce a clear change in what the model wrote.

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

    182) was introduced by Congress in 2011 with the support of President Obama, who in a statement called the legislation “ a common-sense bill that will help ensure that men and women who do equal work receive the equal

  2. #2 peak

    the spellings are inside the whiteboard :)⏎Quality not yet verified by the community.⏎This resource does not contain any images, words or ideas that would upset a reasonable person in any culture.⏎This resource is licensed by LoubeeSav⏎under

  3. #3 peak

    who were tested on the basis of clinical suspicion or risky exposure. A human case of brucellosis was considered confirmed if results of serologic tests, such as ELISA or complement fixation test, were positive, in accordance with the World

  4. #4 peak

    there?⏎There are several types of anxiety disorders. They are:⏎Phobia is about irrational fear of a situation, animal or object, which cannot be explained rationally to the sufferer.⏎A person with a phobia will be triggered by very

  5. #5 peak

    their specifics differ significantly, the dealers can be any wholesale and retail companies that act as intermediaries between the producer and the consumer. If everything is more or less clear with the commodity sphere, then in the service sector, for example, dealers may

  6. #6 peak

    ways in the Balkans. But these unique ecosystems in southeast Europe could soon be gone, along with endangered species such as the Balkan lynx, if plans for over 2,000 dams go ahead, conservationists warn.⏎Western

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 am not ready to learn." In this case, the student is ready to learn. So, "Solving the Problem" is the best way to learn. So, "I am learning in a new environment." So, "

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

    !⏎xs = [6, 9, 13, 42]⏎-- Without library⏎f :: Bool -> [Int]⏎f b = if b then xs else ⏎-- With library⏎g :: Bool -> [Int]⏎g

  2. #2 peak

    while loop:⏎given_arr = [1, 2, 3, 4, 5] i = len(given_arr) - 1 while(i >= 0): print(given_arr[i]) i = i - 1

  3. #3 peak

    > time." (from http://en.wikipedia.org/wiki/Analysis_of_algorithms)⏎> 1 get a positive integer from input⏎> 2 if n> 10⏎> 3 print "This might take a while..."⏎>

  4. #4 peak

    treated as high, and if it is less than 2.5 volts, then it will be treated as low. So,

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. L4:526 excites
  2. L5:1931 suppresses
  3. L5:328 suppresses
  4. L4:361 excites
  5. L6:377 weakly excites
  6. L2:537 weakly excites
  7. L5:1044 weakly excites
  8. L6:2134 weakly excites
  9. L5:2309 weakly suppresses
  10. L5:1792 weakly excites
  11. L3:2323 weakly suppresses
  12. L2:1493 weakly suppresses
  13. L6:1037 weakly suppresses
  14. L2:461 weakly suppresses
  15. L6:195 weakly suppresses
  16. L5:1772 weakly suppresses
  17. L6:792 weakly excites
  18. L0:128 weakly excites
  19. L5:908 weakly suppresses
  20. L1:2868 weakly excites
  21. L5:237 weakly excites
  22. L4:1221 weakly suppresses
  23. L5:399 weakly suppresses
  24. L6:678 weakly excites

Feeds later layers

  1. L12:740 strongly suppresses
  2. L16:2167 suppresses
  3. L10:750 suppresses
  4. L10:659 suppresses
  5. L10:2639 suppresses
  6. L11:2487 suppresses
  7. L11:1725 suppresses
  8. L8:367 excites
  9. L13:2671 suppresses
  10. L13:1286 suppresses
  11. L8:1381 suppresses
  12. L12:249 excites
  13. L11:999 suppresses
  14. L13:2160 suppresses
  15. L8:1920 suppresses
  16. L8:248 suppresses
  17. L13:905 suppresses
  18. L8:691 suppresses
  19. L8:1044 suppresses
  20. L10:545 excites
  21. L13:1508 suppresses
  22. L13:960 excites
  23. L8:158 excites
  24. L10:203 excites

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 -3.57 and -3.421 values between -1.87 and -1.721 values between -1.25 and -1.104 values between -1.10 and -0.945 values between -0.79 and -0.631 values between -0.63 and -0.4819 values between -0.48 and -0.33211 values between -0.33 and -0.171532 values between -0.17 and -0.021782 values between -0.02 and 0.14241 values between 0.14 and 0.2952 values between 0.29 and 0.4524 values between 0.45 and 0.6016 values between 0.60 and 0.765 values between 0.76 and 0.915 values between 0.91 and 1.073 values between 1.07 and 1.221 values between 1.22 and 1.381 values between 1.38 and 1.531 values between 1.53 and 1.691 values between 1.69 and 1.841 values between 2.46 and 2.61offnegativepositiveoften

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. 函
  2. ",-
  3. ={[
  4. acs
  5. atura
  6. 为代表
  7. guard
  8. 媒体
  9. =[⏎
  10. guard

Suppresses

  1. oration
  2. antry
  3. icum
  4. 不认识
  5. IRTH
  6. chưa
  7. syn
  8. itone
  9. ṅ
  10. theros