Neuron Explorer

Qwen3-0.6B

connective before the last list item

A neuron in layer 12 of Qwen3-0.6B. Its strongest snippets are fairly consistent: several of them peak on the same token. It is featured, so it has steering outputs.

What this neuron does

This neuron fires on a word like 'and' or 'then' that comes right before the final item in a list or the last step of a sequence. When it was turned up, the written text did not show a clear change in theme.

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

    forward 2 squares. What have you found?⏎2. 1/4 turn clockwise, forward 2, 1/4 turn anti-clockwise and then forward 3 _______.⏎3. Forward 2, 1/4

  2. #2 peak

    the challenges the Tennessee schools faced.⏎West Virginia, New York, Nebraska, and Delaware also strengthened proficiency standards, while New Mexico, Washington, Hawaii, Montana, and Georgia lowered the bar.⏎Uneven at the Start, a new Education Trust report

  3. #3 peak

    During the last few days we have surveyed butter production costs in 1947 and 1948, compared with 1945 and 194⏎Labor costs in

  4. #4 peak

    these nine orders there exists three ranks: Seraphim, Cherubim, Thrones—Dominions, Authorities, Powers—Principalities, Archangels, and Angels.

  5. #5 peak

    1 chapters. Chapter 1 sets the context for RE and climate change; Chapters 2 through 7 provide information on six RE technologies, and Chapters 8 through 11 address integrative issues. References to chapters and sections are indicated with

  6. #6 peak

    юсь(⏎Y. Раскройте скобки, употребляя глаголы в Present, Past, Future Indefinite, Present, Past, Future Continuous. 1.When I (to open) the window, it

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 things I keep going with, but maybe I could do a little without. What about… if you speak. Maybe not, but let with the length I keep. You can not be. But I am. Anyway. Because I could

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

    isease induced changes in gene flow patterns among Tasmanian devil populations⏎Bruniche-Olsen, A and Burridge, CP and Austin, JJ and Jones, M, Disease induced changes in gene flow patterns among Tasmanian devil populations,

  2. #2 peak

    ; 1945; OFR; 45-37; Fox, I. William; Thigpen, Claude H.; Ginter, Roy L.; Alden, George P.<|endoftext|>Posted September 1st,

  3. #3 peak

    leg ulcer (4) (6) (7).⏎(1) Posnett, J., Gottrup, F., Lundgren, H., Saal, G. The resource impact of wounds on health-care providers in Europe. J W

  4. #4 peak

    if you notice any errors which make this publication unusable.⏎XML: View XML⏎Affleck, David L. R.; Gregoire, Timothy G.; Valentine, Harry T

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. L9:695 excites
  2. L7:94 suppresses
  3. L11:1053 excites
  4. L10:453 suppresses
  5. L11:47 suppresses
  6. L9:770 suppresses
  7. L7:222 weakly suppresses
  8. L7:48 weakly excites
  9. L11:1312 weakly suppresses
  10. L8:1198 weakly excites
  11. L7:75 weakly excites
  12. L9:422 weakly suppresses
  13. L6:354 weakly suppresses
  14. L8:986 weakly excites
  15. L3:674 weakly excites
  16. L10:601 weakly excites
  17. L3:426 weakly excites
  18. L6:1115 weakly suppresses
  19. L10:2952 weakly suppresses
  20. L8:388 weakly suppresses
  21. L7:768 weakly suppresses
  22. L11:1460 weakly suppresses
  23. L9:259 weakly excites
  24. L7:63 weakly excites

Feeds later layers

  1. L13:592 excites
  2. L14:1258 suppresses
  3. L13:2501 excites
  4. L14:683 excites
  5. L13:2789 excites
  6. L25:152 suppresses
  7. L13:1038 weakly excites
  8. L15:1459 weakly excites
  9. L15:2845 weakly excites
  10. L13:2022 weakly excites
  11. L14:858 weakly suppresses
  12. L15:751 weakly suppresses
  13. L19:120 weakly suppresses
  14. L16:1127 weakly excites opening bracket after an imperial measurement
  15. L21:958 weakly excites
  16. L14:776 weakly excites
  17. L14:1382 weakly excites
  18. L13:598 weakly suppresses
  19. L20:2477 weakly suppresses
  20. L18:571 weakly suppresses
  21. L14:2266 weakly excites
  22. L19:228 weakly suppresses
  23. L15:816 weakly excites
  24. L14:720 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.

1 values between -3.19 and -3.061 values between -2.01 and -1.884 values between -1.35 and -1.222 values between -0.96 and -0.835 values between -0.83 and -0.708 values between -0.70 and -0.5615 values between -0.56 and -0.4349 values between -0.43 and -0.30270 values between -0.30 and -0.171045 values between -0.17 and -0.041823 values between -0.04 and 0.09486 values between 0.09 and 0.22100 values between 0.22 and 0.3636 values between 0.36 and 0.4912 values between 0.49 and 0.6211 values between 0.62 and 0.7511 values between 0.75 and 0.8812 values between 0.88 and 1.013 values between 1.01 and 1.144 values between 1.14 and 1.273 values between 1.27 and 1.413 values between 1.41 and 1.541 values between 1.54 and 1.672 values between 1.67 and 1.801 values between 1.93 and 2.06offnegativepositiveoften

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. ocene
  2. bsub
  3. ươ
  4. inct
  5. -has
  6. fixation
  7. pisa
  8. .hasNext
  9. oplast
  10. ntax

Suppresses

  1. mine
  2. rait
  3. gesch
  4. adders
  5. VERR
  6. 出差
  7. 隔壁
  8. 姐姐
  9. scri
  10. arter