Neuron Explorer

Qwen3-0.6B

inside web addresses and citations

A neuron in layer 13 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 is most active inside web addresses and citation codes, such as the "://" in "http://" or the dots in a DOI number. Turning it up made the model switch into Chinese text in several answers, which is a loose link to its strongest negative examples, which are Chinese characters.

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

    . Biological Conservation, 155. 50-58. 10.1016/j.biocon.2012.05.013Full text not available from this repository.⏎Although grass

  2. #2 peak

    JPL. The imaging operations center is based at the Space Science Institute in Boulder, Colo.⏎For more information about the Cassini-Huygens mission visit http://saturn.jpl.nasa.gov/home/index.cfm. The Cassini imaging team

  3. #3 peak

    to supports all kinds of powerful, dynamic applications and services. For more information on this great topic, please consult:⏎- Sun's Java Server Pages home (http://java.sun.com/products/jsp/)⏎- Westy Rockwell's article on "

  4. #4 peak

    receive a test report with each Dataforth SCM. Dataforth test instruments are calibrated and traceable to NIST standards. (see NIST info at: http://www.nist.gov/public_affairs/labs2.htm)View an example of

  5. #5 peak

    8), 1595-1600. http://dx.doi.org/10.1016/j.jbusres.2013.10.001⏎BERISHA-NAM

  6. #6 peak

    Birds Online (T. S. Schulenberg, Editor). Ithaca: Cornell Lab of Ornithology; retrieved from Neotropical Birds Online: http://neotropical.birds.cornell.edu/

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 can be something else, but I can be something else... ... I can be something... ( ... ) ... I can be something... ( ... ) ... I ... ( ... ) ... I can be something ... ( ... ) ...

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

    ore 2-cleft to base.⏎One species: Mediterranean region; cultivated and adventive worldwide, including in China.⏎(Authors: She Menglan (佘孟兰 Sheh Meng-lan); Mark F. Watson)<|endoftext|>By Cary Gron

  2. #2 peak

    on Chinese Auxiliary Verbs for Thai Learners Based on the Comparative Analysis⏎- 人文学院-已发表论文 ⏎“助动词”在汉语中具有较大的作用,使用率高,但是泰国学生经常在学习和使用汉语

  3. #3 peak

    's longest sustained dynasties. Although it was founded by King Park Hyeokgeose, the dynasty was ruled by the Gyeongju Kim (김, 金) clan for most of its 992-year history.

  4. #4 peak

    heating and air conditioning systems, it also means that pollutants that originate indoors are getting trapped inside.⏎“People normally<|endoftext|>Spring Dawn by Meng Haoran 春眠不觉晓 处处闻啼茑 夜来风雨声 花

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. L10:155 strongly excites
  2. L8:328 strongly excites
  3. L12:158 strongly suppresses
  4. L10:669 strongly suppresses
  5. L4:76 strongly excites
  6. L11:223 strongly excites
  7. L2:24 strongly suppresses
  8. L7:92 strongly excites
  9. L5:7 strongly suppresses
  10. L9:17 strongly suppresses
  11. L4:342 suppresses
  12. L1:25 excites
  13. L1:699 excites
  14. L10:1388 suppresses
  15. L11:416 excites
  16. L3:1256 excites
  17. L0:83 suppresses
  18. L3:130 excites
  19. L9:74 excites
  20. L4:398 suppresses
  21. L7:72 excites
  22. L9:1722 suppresses
  23. L10:39 suppresses
  24. L3:14 excites

Feeds later layers

  1. L15:59 strongly excites
  2. L15:184 strongly suppresses
  3. L17:611 strongly suppresses
  4. L15:390 strongly suppresses
  5. L18:66 strongly excites
  6. L17:56 strongly excites
  7. L14:744 strongly suppresses
  8. L16:513 strongly suppresses
  9. L17:1016 strongly suppresses
  10. L23:59 strongly suppresses
  11. L17:812 strongly suppresses
  12. L17:473 strongly suppresses
  13. L17:257 strongly suppresses
  14. L27:54 strongly suppresses
  15. L26:81 strongly suppresses
  16. L16:899 strongly suppresses
  17. L15:1033 strongly suppresses
  18. L16:771 strongly suppresses
  19. L19:68 strongly suppresses full stops closing a claim
  20. L17:98 strongly suppresses
  21. L17:535 strongly suppresses
  22. L27:269 strongly suppresses
  23. L19:119 strongly excites
  24. L21:279 strongly 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 -1.72 and -1.643 values between -1.48 and -1.401 values between -1.32 and -1.233 values between -1.23 and -1.152 values between -1.15 and -1.072 values between -1.07 and -0.994 values between -0.99 and -0.914 values between -0.91 and -0.8310 values between -0.83 and -0.757 values between -0.75 and -0.6612 values between -0.66 and -0.589 values between -0.58 and -0.5013 values between -0.50 and -0.4219 values between -0.42 and -0.3421 values between -0.34 and -0.2630 values between -0.26 and -0.1844 values between -0.18 and -0.0951 values between -0.09 and -0.01104 values between -0.01 and 0.0765 values between 0.07 and 0.1574 values between 0.15 and 0.23156 values between 0.23 and 0.31291 values between 0.31 and 0.40519 values between 0.40 and 0.48722 values between 0.48 and 0.56720 values between 0.56 and 0.64512 values between 0.64 and 0.72323 values between 0.72 and 0.80136 values between 0.80 and 0.8839 values between 0.88 and 0.978 values between 0.97 and 1.052 values between 1.05 and 1.131 values between 1.45 and 1.54offnegativepositiveoften

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. three
  2. four
  3. second
  4. seven
  5. five
  6. reason
  7. ten
  8. balance
  9. lot
  10. profit

Suppresses

  1. 分
  2. 在
  3. 考虑
  4. 以
  5. 是在
  6. 组织
  7. 安全
  8. 正
  9. 包含
  10. 都