Neuron Explorer

Qwen3-0.6B

comma after an introductory word like then or when

A neuron in layer 6 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 that comes right after a short opening word such as "Then", "When", "How" or "If", where a clause is about to be set apart. Turning it up gave text with no obvious link to the comma pattern.

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

    McIntire started out on his journey westward not knowing just where Fortune would lead him. He spent almost a year in Missouri, driving a stage coach. Then, in company with three companions, he drove a herd of sheep from Missouri to central

  2. #2 peak

    that take this arithmetic thinking into account or, more recently, developing approaches to elementary school mathematics that build foundations of algebraic reasoning

  3. #3 peak

    remains poorly understood,” the Editors say. “This has led to large uncertainties in modelled estimates of terrestrial carbon storage and carbon cycle–climate feedbacks.” How, then, can important measures with such large uncertainties provide any confidence about temperature predictions for

  4. #4 peak

    .⏎We were confined three days and nights⏎Within those iron bars,⏎When, oh! the welcome news then came,⏎To get abound the cars.⏎They marched

  5. #5 peak

    can use some, or all, of this content as you see fit, whether that’s to compliment, vary or boost work set by your child's school or, of course, just to give some extra structure to your usual EdPlace approach.⏎

  6. #6 peak

    an epileptic," but rather "a person with epilepsy."⏎When writing a story or advertisement, the writer should use the term "people with disabilities" exclusively or, at a minimum, as the initial reference. Subsequent references can us terms like

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 problem of a person who is not in a good position to live. But I think that this is not the same as being a good person. So, I am thinking about a person who is not in a good position to live,

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

    article, Brian Sekita discusses how GIS is vital to all phases of the emergency management process.⏎Natural disasters such as hurricanes, floods, tornadoes, earthquakes, and volcanic eruptions have impacted the globe since the beginning of time. Technological improvements

  2. #2 peak

    follow.⏎But heatwaves do happen anyway. To find out if the current heatwave was made more likely by climate change, a team at World Weather Attribution led by Friederike Otto of the University of Oxford, UK has conducted a rapid-response

  3. #3 peak

    ase.⏎The endonuclease enables for the virus to trick someone into producing viral proteins, which join into new viruses that spread the infection.⏎The team led by EMBL head Stephen Cusack analyzed crystals of endonuclease from

  4. #4 peak

    (ICTs) can mitigate many issues in SIDS and help to diversify their economies.⏎She added that innovations such as artificial intelligence, blockchain, drones, and mobile money were enabling progress towards sustainable development in SIDS, but this is constrained

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. L5:694 strongly excites
  2. L5:1821 excites
  3. L5:395 suppresses
  4. L3:373 suppresses
  5. L2:89 suppresses
  6. L5:351 suppresses
  7. L2:180 suppresses 'is' after focus or goal nouns
  8. L5:2747 weakly excites
  9. L4:12 weakly suppresses
  10. L2:2124 weakly suppresses
  11. L5:995 weakly excites
  12. L5:99 weakly excites
  13. L3:552 weakly excites
  14. L5:781 weakly suppresses
  15. L5:1217 weakly excites
  16. L3:466 weakly excites
  17. L5:976 weakly excites
  18. L2:1394 weakly excites
  19. L4:1587 weakly excites
  20. L3:700 weakly suppresses
  21. L5:2643 weakly suppresses
  22. L2:24 weakly excites
  23. L4:139 weakly suppresses
  24. L2:1185 weakly suppresses

Feeds later layers

  1. L7:288 excites
  2. L10:142 excites opening parenthesis mid-sentence
  3. L14:220 excites
  4. L17:2787 excites comma after 'that' or a verb before an aside
  5. L13:1721 suppresses
  6. L16:2571 excites
  7. L9:819 excites
  8. L9:1371 excites
  9. L17:2597 excites
  10. L15:1213 excites
  11. L16:1767 excites
  12. L14:774 suppresses
  13. L14:2402 excites
  14. L19:1864 excites
  15. L16:1253 excites
  16. L13:897 suppresses
  17. L15:770 excites
  18. L11:1272 excites unclear: sentence ends in factual prose
  19. L13:685 weakly suppresses
  20. L17:2084 weakly suppresses
  21. L21:2772 weakly excites
  22. L20:1466 weakly excites
  23. L8:204 weakly suppresses
  24. L13:2609 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.

2 values between -2.67 and -2.491 values between -2.30 and -2.111 values between -1.93 and -1.743 values between -1.74 and -1.552 values between -1.55 and -1.373 values between -1.37 and -1.186 values between -1.18 and -0.998 values between -0.99 and -0.818 values between -0.81 and -0.6234 values between -0.62 and -0.4398 values between -0.43 and -0.25703 values between -0.25 and -0.062422 values between -0.06 and 0.13498 values between 0.13 and 0.3158 values between 0.31 and 0.5020 values between 0.50 and 0.6915 values between 0.69 and 0.8710 values between 0.87 and 1.062 values between 1.06 and 1.254 values between 1.25 and 1.432 values between 1.43 and 1.621 values between 1.62 and 1.811 values between 1.81 and 1.991 values between 1.99 and 2.181 values between 2.55 and 2.742 values between 2.74 and 2.921 values between 3.48 and 3.671 values between 4.60 and 4.79offnegativepositiveoften

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. shaw
  2. __,
  3. ,**
  4. 为主体
  5. rows
  6. **,
  7. *,
  8. awning
  9. 为主的
  10. -,

Suppresses

  1. opping
  2. cg
  3. ORM
  4. ARENT
  5. slick
  6. ogonal
  7. 功课
  8. cg
  9. budget
  10. cj