Neuron Explorer

Qwen3-0.6B

comma inside a question

A neuron in layer 20 of Qwen3-0.6B. Its strongest snippets are very consistent: they mostly peak on the same token. It is featured, so it has steering outputs.

What this neuron does

This neuron fires on a comma in the middle of a question, usually right before an alternative such as "or" (for example "Are statements descriptive, prescriptive or emotive,"). Its opposite is the word just before a question mark, like "later?" or "right?". Turning it up or down made the model write strings of question marks or odd symbols.

Written by Claude from this neuron's activations and steering outputs. Confidence: medium. 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

    is the sequence of events in the conversion of the chemical energy derived from ATP hydrolysis to a mechanical stroke? How are these altered in a processive motor, where each of the catalytic heads is subject to mechanical forces from intramolecular

  2. #2 peak

    -based learning can create developmental imbalances we never anticipated.⏎Are you really going to read this blog word for word? Or are you much more likely to skim it, looking for ideas that, for whatever reason, resonate with you?⏎Will you resist

  3. #3 peak

    use of ethical language.⏎What is the meaning and nature of a moral judgement? Are statements descriptive, prescriptive or emotive, or do they correspond to an objective, universal goodness which we attempt to access through

  4. #4 peak

    they have refused to host a public hearing. Are they afraid that they would again have a standing-room-only crowd of concerned citizens rejecting the HCP, as they did in 2016? Are they trying to avoid having

  5. #5 peak

    one interesting class of question is aimed at finding out the variety of situations those laws allow. For example, is there only a single kind of rotating black hole, or do the laws of general relativity admit an infinite variety of such objects?

  6. #6 peak

    ’t carefully considered before. Is it someone with an advanced degree? Is it a Nobel Prize winner? Is it someone who has memorized a great deal of information, or someone who knows how to find and evaluate information? A biological researcher? A

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: Yes? No?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?

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

    originated in Europe 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

  2. #2 peak

    place.⏎Ready to start? Let’s dive in!⏎A little history of the SSL certificate⏎They say “to know your future you must know your past”, right? Put into a closer, business perspective, I’d rephrase that with “

  3. #3 peak

    ’s Heartland Theory is, why this theory is important and what the strengths and limitations of this theory are.⏎Ready to learn more about Mackinder’s Heartland Theory? Read on…⏎- What is Mackinder’s Heartland Theory?⏎- Why

  4. #4 peak

    .”⏎SALLY FIELD, acceptance speech at 2007 Emmy Awards⏎What differences do you see in countries who have endured war compared to those who have not?⏎Featured image: children during Pol Pot provided by http://www.ridec

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. L16:1078 strongly excites
  2. L18:2812 strongly excites
  3. L17:1069 strongly excites
  4. L16:2817 strongly suppresses
  5. L15:2412 strongly suppresses
  6. L11:296 suppresses
  7. L17:2579 suppresses
  8. L15:1473 excites
  9. L5:1800 suppresses
  10. L16:1123 excites
  11. L17:1407 suppresses "and" joining questions (wh-question context)
  12. L13:2080 suppresses
  13. L13:2939 excites
  14. L16:669 excites
  15. L6:1476 excites
  16. L5:458 excites
  17. L14:2652 weakly excites
  18. L13:943 weakly excites
  19. L2:1640 weakly excites
  20. L1:2000 weakly excites
  21. L15:725 weakly suppresses
  22. L14:1800 weakly suppresses
  23. L13:2701 weakly excites
  24. L17:142 weakly suppresses

Feeds later layers

  1. L24:402 strongly suppresses
  2. L25:541 strongly suppresses
  3. L27:233 strongly suppresses
  4. L26:613 strongly suppresses
  5. L23:164 strongly suppresses
  6. L21:1499 strongly suppresses question marks ending a question
  7. L26:953 strongly suppresses
  8. L27:838 suppresses
  9. L27:975 suppresses
  10. L24:2628 suppresses
  11. L27:1576 excites
  12. L27:1080 excites
  13. L21:1538 suppresses
  14. L24:2138 suppresses comma before 'right?' tag
  15. L26:1767 suppresses
  16. L25:1025 suppresses
  17. L22:321 weakly excites
  18. L23:2365 weakly excites
  19. L23:100 weakly excites comma or full stop ending a clause
  20. L26:775 weakly suppresses
  21. L27:1551 weakly excites
  22. L23:1526 weakly suppresses
  23. L24:1922 weakly excites
  24. L24:902 weakly 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.

3 values between -13.01 and -12.012 values between -12.01 and -11.015 values between -11.01 and -10.024 values between -10.02 and -9.023 values between -9.02 and -8.025 values between -8.02 and -7.0214 values between -7.02 and -6.039 values between -6.03 and -5.0319 values between -5.03 and -4.0325 values between -4.03 and -3.0340 values between -3.03 and -2.0492 values between -2.04 and -1.04985 values between -1.04 and -0.042579 values between -0.04 and 0.9677 values between 0.96 and 1.9511 values between 1.95 and 2.956 values between 2.95 and 3.952 values between 3.95 and 4.954 values between 4.95 and 5.947 values between 5.94 and 6.941 values between 6.94 and 7.942 values between 7.94 and 8.942 values between 8.94 and 9.933 values between 10.93 and 11.933 values between 11.93 and 12.931 values between 12.93 and 13.921 values between 13.92 and 14.922 values between 14.92 and 15.921 values between 25.89 and 26.89offnegativepositiveoften

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. ':{'
  2. jerne
  3. //--------------------------------------------------------------⏎
  4. ':['
  5. >()->
  6. )::
  7. ":-
  8. regation
  9. ',[
  10. ⾦

Suppresses

  1. ?
  2. ?
  3. ?⏎⏎
  4. ?⏎⏎
  5. ?⏎
  6. ?⏎
  7. ?"
  8. ? ⏎
  9. ?”
  10. ?”