Neuron Explorer

Qwen3-0.6B

comma after 'though' mid-sentence

A neuron in layer 5 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 the comma that comes right after the word 'though' when it sits in the middle of a sentence, as in 'the main city, though, ...'. Turning it up did not produce comma-heavy text; the output wandered into unrelated topics.

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

    start of the 1st century A.D., when the main city of Populonia was mysteriously abandoned. People continued to live at the villa, though, until the end of the 5th century, so it is one of the

  2. #2 peak

    the curves, though, when your spine is seen from the rear, it looks like it’s running straight

  3. #3 peak

    called ruby-throats, since the males share the same red gorget as their East Coast cousins. The predominant species of hummingbirds in our valley, though, is the broad-tail. These birds look similar to the ruby-throats

  4. #4 peak

    underestimated, an error that has in effect enabled others to have a stronger claim on their wealth creation role.⏎It is hard to make the pitch for government, though,

  5. #5 peak

    whether the health benefits of chocolate outweigh the additional calories consumed. Don’t be concerned about eating the occasional chocolate bar. If you are trying to lose weight, though, eating more fruits and vegetables is a better option.<|endoftext|>MLA style for documentation

  6. #6 peak

    habitat, given the unpredictable nature of the cone crop. Forests that also have at least some areas of pines, Norway Sprue or Larch, though, can sustain large populations. Some Red Squirrels do fall prey to Pine Mart

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 using this website to find the most recent of the week: I have been told by the teacher of this week: I have been given this to you: I have been given this to you: I have been given this to

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

    assets) depending on which way the decision goes.⏎Those are the ‘easy’ situations<|endoftext|>You hear it all the time, “don’t write your picture book in rhyme.” That’s because

  2. #2 peak

    awesome) discovery. Without knowing too much about it, I’m sure that there will be a raft of problems involving

  3. #3 peak

    alin. Amygdalin is a chemical ingredient found in the pits of many fruits, raw nuts, and plants (see Question 1).⏎- Hydrogen cyanide is thought to be the main anticancer compound

  4. #4 peak

    “, and others in my awesome Edumatch Voxer group, told me about BreakOut EDU which is basically the same thing as an Escape Room. Best of all, they have resources, templates, and ready-made games which you or your

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. L0:2311 weakly suppresses
  2. L4:813 weakly suppresses
  3. L1:2502 weakly excites
  4. L1:1693 weakly suppresses
  5. L1:969 weakly excites
  6. L2:1996 weakly suppresses
  7. L3:2664 weakly suppresses
  8. L2:219 weakly excites
  9. L1:1712 weakly excites
  10. L1:1331 weakly suppresses
  11. L1:2778 weakly excites
  12. L1:1559 weakly excites
  13. L0:319 weakly suppresses blank space after a label colon
  14. L4:2989 weakly suppresses
  15. L2:1175 weakly excites
  16. L3:369 weakly excites
  17. L3:1966 weakly excites
  18. L1:188 weakly suppresses
  19. L4:2357 weakly excites
  20. L4:594 weakly suppresses
  21. L1:2646 weakly suppresses
  22. L3:203 weakly excites
  23. L3:1262 weakly excites
  24. L1:1992 weakly suppresses

Feeds later layers

  1. L14:955 weakly suppresses
  2. L16:2988 weakly excites
  3. L25:856 weakly excites
  4. L19:1431 weakly suppresses
  5. L13:1355 weakly excites
  6. L6:70 weakly excites
  7. L23:1666 weakly suppresses
  8. L20:1826 weakly suppresses
  9. L8:781 weakly suppresses
  10. L8:299 weakly suppresses
  11. L12:2276 weakly excites
  12. L14:134 weakly suppresses spot right before a citation year
  13. L12:2425 weakly suppresses
  14. L14:647 weakly excites
  15. L10:838 weakly suppresses comma in a list of paired nouns
  16. L25:66 weakly excites
  17. L25:2406 weakly suppresses
  18. L19:369 weakly excites
  19. L7:2608 weakly suppresses
  20. L13:161 weakly suppresses
  21. L10:2712 weakly excites
  22. L26:1969 weakly excites
  23. L11:56 weakly excites
  24. L6:1073 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 -0.83 and -0.751 values between -0.75 and -0.671 values between -0.67 and -0.594 values between -0.59 and -0.516 values between -0.51 and -0.4329 values between -0.43 and -0.3563 values between -0.35 and -0.27169 values between -0.27 and -0.19379 values between -0.19 and -0.11739 values between -0.11 and -0.031393 values between -0.03 and 0.05644 values between 0.05 and 0.13280 values between 0.13 and 0.21113 values between 0.21 and 0.2954 values between 0.29 and 0.379 values between 0.37 and 0.458 values between 0.45 and 0.536 values between 0.53 and 0.613 values between 0.61 and 0.691 values between 0.69 and 0.771 values between 0.77 and 0.851 values between 0.85 and 0.921 values between 1.00 and 1.081 values between 1.64 and 1.721 values between 2.28 and 2.36offnegativepositiveoften

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. uster
  2. meet
  3. toi
  4. gen
  5. 胆
  6. ánh
  7. widths
  8. get
  9. issues
  10. args

Suppresses

  1. candidacy
  2. 的时间里
  3. MARY
  4. 儿
  5. 了自己的
  6. eman
  7. 兒
  8. ikal
  9. 写下
  10. 了过来