Neuron Explorer

Qwen3-0.6B

commas between numbers in lists

A neuron in layer 27 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 sits between numbers in a list or sequence, such as '1,2,3,4' or '143,754', and to a lesser extent on commas between items in a list. Its negative side is weak and comes from ordinary commas in prose. Turning it up made the text repeat words like 'much', and turning it down produced repeated Chinese characters, so the steering results do not clearly reflect the 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

    still useful. Arabic numerals share their origin from the numbering system of India.⏎The pattern of Arabic numerals start from 0,1,2,3,4 and it simply goes up in such way. On the other hand Roman numer

  2. #2 peak

    icata|⏎|Species Authority:||Todd, 1913|⏎|Red List Category & Criteria:||Endangered B1ab(i,ii,iii,v) ver 3.1|⏎|Reviewer(s):||Butchart

  3. #3 peak

    ,182; Aurora, 286,028; Lakewood, 143,754; Fort Collins, 124,665;

  4. #4 peak

    link="none" columns="2" size="medium" ids="2568,2567,2569,2570"]These

  5. #5 peak

    couch or the bed. Shelter,food,clothes,education, and family are all the things that every person deserves. These are all basic needs

  6. #6 peak

    Previous: [SVG] How to draw a rectangle?⏎<svg height="210" width="500"> <polygon points="200,10 250,190 160,2

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:并并并并并并并并并并并并并并并并并并并并并并并并并并并并并并并并并并并并并并并并并并并并并并并并

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

    about how to correct material in RePEc.⏎For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: (Michael McNulty)⏎If references are entirely missing, you can add

  2. #2 peak

    提要 [ summary or outline of a book, play, etc ]⏎note: 字根組合的概念是 "general view",可以用來指電視劇、書<|endoftext|>Making the decision for Google to change google.ps’ website to mention “

  3. #3 peak

    法、分析法来进行研究,探讨泰国学生在学习汉语助动词中的困难及偏误情况,分析他们产生偏误的原因。我们发现经过基于对比研究设计的汉语

  4. #4 peak

    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助动词中产生各种的

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. L25:535 strongly excites commas inside number lists
  2. L25:1757 strongly suppresses
  3. L26:1876 excites
  4. L20:1298 suppresses
  5. L20:1385 suppresses
  6. L15:184 excites
  7. L19:43 suppresses
  8. L19:1826 suppresses
  9. L25:422 excites
  10. L18:251 excites
  11. L26:75 suppresses
  12. L10:155 excites
  13. L18:66 suppresses
  14. L13:44 suppresses inside web addresses and citations
  15. L24:248 excites
  16. L4:3 excites commas after citation fields
  17. L16:7 excites
  18. L2:422 suppresses
  19. L16:240 excites comma after 'sometimes' or 'the truth'
  20. L5:97 excites
  21. L14:35 suppresses
  22. L25:231 weakly excites
  23. L0:132 weakly excites
  24. L18:2403 weakly excites

Feeds later layers

None: this is the last layer.

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.

847 values between -2.12 and -0.143048 values between -0.14 and 1.845 values between 1.84 and 3.821 values between 3.82 and 5.803 values between 5.80 and 7.781 values between 7.78 and 9.762 values between 9.76 and 11.741 values between 75.08 and 77.06offnegativepositiveoften

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. inals
  2. .ravel
  3. Pieces
  4. (IN
  5. ائية
  6. ians
  7. .flatten
  8. agger
  9. eldorf
  10. much

Suppresses

  1. 并
  2. 则
  3. 并且
  4. 并将
  5. 请
  6. 并对
  7. 在
  8. 并通过
  9. 请你
  10. 并于