Neuron Explorer

Qwen3-0.6B

unclear: mixed, repeated phrasing

A neuron in layer 14 of Qwen3-0.6B. Its strongest snippets are mixed: they peak on different tokens. It is featured, so it has steering outputs.

What this neuron does

This neuron fires on many unrelated spots in text, such as a comma in a list, the word 'characterized' after 'disorder', and the start of a repeated phrase. No single pattern stands out, so the meaning is unclear.

Written by Claude from this neuron's activations and steering outputs. Confidence: low. 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

    physical ability to perform the movements. (Source: excerpt from NINDS Apraxia Information Page: NINDS)⏎Apraxia is a neurological disorder characterized by loss of the ability to execute... (Source: excerpt from NINDS

  2. #2 peak

    eye that sees, a hand that obeys, and a soul that feels.” Through art activities, children learn how to “Our schools prepare an eye that sees, a hand that obeys, and a soul that feels

  3. #3 peak

    of electromagnetic energy. They move through space at the speed of light and can be man-made or occur naturally. They move through space at the speed of light and can be man-made or occur naturally

  4. #4 peak

    . One out of three people aged 65-74 has some level of hearing loss. After age 75, that ratio goes up to one out of every two people.⏎Researchers don't fully understand why hearing decreases with

  5. #5 peak

    33.zip⏎It will create a folder as /var/www/html/drupal-7.33, I will rename the folder as /var/www/html/drupal.⏎# mv /var/www/html/drupal-7.33/

  6. #6 peak

    and use them to print output at certain stages of the program. For example:⏎>>> >>> s1 = 'String in single quotes' >>> s1 'String in single quotes' >>> >>> s2 = "String in double quotes" >>> s

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 number of these things, the number of things, the number of things, and the things, the things, and the things, and the things, the things, and the things, and the things, the things, and the things

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

    new sentences. Suddenly, the genetic equivalent of “The rain in Spain stays mainly on the plain” becomes “Mainly the rain in Spain stays mainly on the plain” with the word “mainly” acting as a retrotransposon.

  2. #2 peak

    this word list.⏎Words formed by adding one letter before or after endoplasm (in bold), or to adelmnops in any order:⏎s - endoplasms⏎All words formed from endoplasm by changing one letter⏎Browse

  3. #3 peak

    arkshire and King William I granted them Arbuthnott in Mearns.⏎Spelling variations of this family name include: Oliphant, Olifant, Olifard and others.⏎First found in Perthshire (Gaelic

  4. #4 peak

    difficult because she can't just repeat what the teacher says. She has to swap the words. "I give you the brush" becomes "you give me the brush.") The two main things she hasn't learned at all are tense (was/is

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. L13:1208 excites
  2. L13:1745 excites
  3. L12:849 excites
  4. L13:2214 suppresses
  5. L13:3055 excites
  6. L11:893 excites
  7. L13:1749 suppresses
  8. L13:411 excites
  9. L13:51 excites
  10. L12:371 suppresses
  11. L13:2442 weakly suppresses
  12. L9:964 weakly suppresses
  13. L13:1524 weakly suppresses
  14. L13:2664 weakly excites
  15. L11:1848 weakly excites
  16. L9:409 weakly suppresses
  17. L13:2086 weakly excites
  18. L13:982 weakly suppresses
  19. L12:1071 weakly suppresses
  20. L13:1993 weakly suppresses
  21. L13:2309 weakly suppresses
  22. L11:352 weakly suppresses
  23. L6:1919 weakly excites
  24. L12:15 weakly suppresses

Feeds later layers

  1. L17:2570 excites
  2. L15:1357 excites
  3. L16:2261 excites
  4. L16:998 excites
  5. L16:2075 excites
  6. L16:500 excites
  7. L15:2154 suppresses
  8. L16:2882 suppresses
  9. L15:1994 excites
  10. L16:1601 suppresses
  11. L15:2089 excites
  12. L17:1976 excites
  13. L17:2846 excites
  14. L15:2218 suppresses
  15. L15:1435 weakly excites
  16. L17:1421 weakly excites
  17. L15:1892 weakly excites
  18. L16:1512 weakly excites
  19. L15:2966 weakly suppresses
  20. L16:2806 weakly suppresses
  21. L16:2717 weakly suppresses
  22. L23:538 weakly suppresses
  23. L18:981 weakly suppresses
  24. L15:1628 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.

2 values between -1.10 and -0.971 values between -0.84 and -0.728 values between -0.72 and -0.5948 values between -0.59 and -0.46228 values between -0.46 and -0.34787 values between -0.34 and -0.211300 values between -0.21 and -0.081106 values between -0.08 and 0.04331 values between 0.04 and 0.1759 values between 0.17 and 0.3016 values between 0.30 and 0.428 values between 0.42 and 0.554 values between 0.55 and 0.674 values between 0.67 and 0.802 values between 0.80 and 0.931 values between 1.05 and 1.181 values between 1.43 and 1.561 values between 2.95 and 3.081 values between 3.84 and 3.96offnegativepositiveoften

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. zte
  2. fühl
  3. instead
  4. acers
  5. getTable
  6. ylan
  7. coração
  8. ANGED
  9. aleb
  10. qli

Suppresses

  1. EMBER
  2. ainer
  3. ething
  4. aturity
  5. dition
  6. ăng
  7. طبع
  8. 植
  9. 文本
  10. text