Neuron Explorer

Qwen3-0.6B

mid-word fragments of "cognitive" and similar words

A neuron in layer 3 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 the middle piece of a word that the model has split into parts, most strongly in "cognitivists" and "cognitively". Other examples are fragments in "scholarworks" and "snarling". The pattern is not very consistent beyond that, and turning it up did not give a 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

    for the greatest number (Utilitarianism), it can always be questioned.⏎People can still have differences of opinion about things.⏎“Ethical naturalists are cognitivists, who believe that moral statements are propositional, and so can be

  2. #2 peak

    als hunted birds for their fashionable feathers for thousands of years and may have exploited certain plants for their medicinal properties–compelling evidence that our hominin cousins were cognitively sophisticated.⏎-<|endoftext|>During the

  3. #3 peak

    a test, so fearful of getting the wrong answers that they do not complete as many items as their peers. But test anxiety is actually quite different from a cognitively-based slow processing speed.⏎However, many observations of anxious kids suggest that the

  4. #4 peak

    ched body, weight shifted away from the fear-eliciting stimulus) accompanied by aggressive signs such as piloerection, barking, growling, snarling and biting. Dilated pupils, increased respiratory rate and a rapid heart rate

  5. #5 peak

    ⏎To perform the test, the patient is instructed to stand on one leg without support of the upper extremities or bracing of the unweighted leg against the stance leg. The patient begins the test with the eyes open, practicing once or twice

  6. #6 peak

    or emotive, or do they correspond to an objective, universal goodness which we attempt to access through moral debates? What does ‘good’ mean?⏎Cognitivists – moral statements are propositional and therefore deal with verifiable facts⏎

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 given to that. I have been given to that. I have been given to that. I have been given to that. I have been given to that. I have been given to that. I have been given to that

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

    , honey and bees, fruits, and the like; and uses its fore paws with a deftness almost equal to that of a monkey. It is found from Central Mexico to

  2. #2 peak

    White Fragility: Why It’s So Hard for White People to Talk About Racism by Robin Diangelo⏎“Antiracist educator Robin DiAngelo deftly illuminates the phenomenon of white fragility and “allows us to understand racism

  3. #3 peak

    meetings. They were prepared to award her the prize but Houdini protested that he needed to see for himself. At the seance, Houdini saw through the deception and called “Margery” on it, much to the annoyance of

  4. #4 peak

    attached to the museum's awkward juxtaposition with the state Division of Taxation building.⏎Beginning in 1973, the N.J. Supreme Court saw fit to interject itself into education policy and

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. L2:1321 weakly suppresses
  2. L1:1476 weakly suppresses
  3. L2:1888 weakly excites
  4. L2:2292 weakly excites
  5. L1:2230 weakly suppresses
  6. L1:2186 weakly excites
  7. L0:2071 weakly excites
  8. L0:1082 weakly suppresses
  9. L0:1297 weakly suppresses
  10. L2:463 weakly excites
  11. L0:1941 weakly suppresses
  12. L1:208 weakly suppresses
  13. L2:715 weakly suppresses the word 'following' after 'the'
  14. L0:777 weakly excites
  15. L0:3020 weakly excites
  16. L2:2810 weakly suppresses
  17. L1:702 weakly excites
  18. L2:1178 weakly excites
  19. L1:2543 weakly suppresses
  20. L2:2119 weakly suppresses
  21. L2:1356 weakly suppresses
  22. L2:2367 weakly suppresses
  23. L0:1413 weakly excites
  24. L1:2915 weakly suppresses

Feeds later layers

  1. L4:285 suppresses
  2. L16:836 weakly excites
  3. L10:1431 weakly excites
  4. L4:3046 weakly excites
  5. L4:3071 weakly excites
  6. L4:2893 weakly excites the word 'means' after a subject
  7. L4:2976 weakly excites
  8. L20:1856 weakly suppresses
  9. L18:2934 weakly suppresses
  10. L19:531 weakly excites
  11. L6:616 weakly excites
  12. L6:1906 weakly suppresses
  13. L19:1967 weakly excites
  14. L6:165 weakly suppresses
  15. L22:2504 weakly suppresses
  16. L19:513 weakly suppresses
  17. L15:2944 weakly suppresses
  18. L20:1907 weakly suppresses
  19. L4:1444 weakly suppresses
  20. L4:2715 weakly excites
  21. L10:2075 weakly excites
  22. L17:1619 weakly excites
  23. L17:1987 weakly excites
  24. L8:2421 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 -7.70 and -7.431 values between -2.65 and -2.391 values between -1.86 and -1.593 values between -1.32 and -1.064 values between -1.06 and -0.7911 values between -0.79 and -0.5349 values between -0.53 and -0.261834 values between -0.26 and 0.001927 values between 0.00 and 0.2763 values between 0.27 and 0.545 values between 0.54 and 0.804 values between 0.80 and 1.072 values between 1.07 and 1.331 values between 1.33 and 1.601 values between 2.13 and 2.391 values between 2.66 and 2.92offnegativepositiveoften

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. olia
  2. XD
  3. unch
  4. 吉利
  5. beiter
  6. вел
  7. LowerCase
  8. ghan
  9. xed
  10. REPL

Suppresses

  1. handleError
  2. liği
  3. Matter
  4. .handler
  5. 一对
  6. handler
  7. handled
  8. eries
  9. cbc
  10. iner