Neuron Explorer

Qwen3-0.6B

"of" after a noun phrase

A neuron in layer 8 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 responds weakly to the word "of" when it follows a noun, as in "a dictionary of", "center of attention" or "pattern of events". Its negative side is unclear, and the steering results did not show a clear theme.

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

    review previous lesson, summarize periodically).⏎- Allow time for clarification of directions and essential information.⏎- Provide study guides or review sheets for exams.<|endoftext|>A modern dictionary of Catholic terms, both common and obscure. Find accurate definitions of words and phrases.⏎

  2. #2 peak

    . Grant Wood AEA empowers each learner with the accessibility and assistive technologies that ensure a culture of inclusivity, equity and improve student achievement and outcomes.⏎Want to get the conversation

  3. #3 peak

    other organization and deliver them to those in need. The individuals will have a smile on their face seeing your little one and he/she will love to be the center of attention on the trip!⏎If you are involved with your child’s school, make

  4. #4 peak

    elligent Design holds that design is empirically detectable in nature, and particularly in living systems.⏎A. What is a "design?" A Design is a Pattern of Events Arranged by Intent.⏎A design (noun) is nothing more than a

  5. #5 peak

    , i.e. the belief that there is but one God. As a matter of fact Monotheism (TOWHID) is not only a principle of the religion, but the most important of the tenets. It is the very

  6. #6 peak

    and learn from them.⏎There are many books about fashion. Here are a few books that will help you understand fashion:⏎- Fashion: The Definitive History of Costume and Style⏎- Stylist: The Interpreters of Fashion⏎-

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: of course, the idea of the people of the world, of course, we need to think about the nature of the problem that the people of the world, of course, they are now more likely to be found to be in the same

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

    future careers. Please be sure to visit www.LL-inc.org for activities to facilitate career-conversation with your young leader.<|endoftext|>How does Utterson perceive the relationship between Jekyll and Hyde for most of the novel? Is his interpretation understandable?

  2. #2 peak

    and its impacts on individual and social wellbeing. We draw on examples to synthesise existing research, analyse concepts and surface the societal and ethical challenges arising from the blurring of the boundaries of health data. We chart the mechanisms through which datafication

  3. #3 peak

    , colonization of multiple urban centres will be particularly difficult when urbanization requires genetic adaptations, having implications for urban species diversity.⏎The rapid growth in urban development and increasing concentration of people in these areas (Cohen 2003; UN

  4. #4 peak

    .⏎Studying public opinion in Japan and South Korea is critically important for exploring the causes and consequences of the deterioration of the relationship between the two countries. Japan–South Korea relations are at their worst level since

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. L1:911 weakly excites
  2. L6:1009 weakly excites
  3. L5:940 weakly suppresses
  4. L7:2611 weakly excites
  5. L6:1810 weakly excites
  6. L7:806 weakly suppresses
  7. L5:20 weakly suppresses
  8. L4:1076 weakly excites
  9. L7:2264 weakly suppresses
  10. L6:2862 weakly suppresses
  11. L4:944 weakly suppresses
  12. L2:2821 weakly excites
  13. L6:237 weakly excites
  14. L6:1258 weakly excites "mean by" in questions and definitions
  15. L5:1271 weakly excites
  16. L7:224 weakly suppresses
  17. L7:1316 weakly suppresses
  18. L5:2980 weakly excites 'host' or 'number' before 'of'
  19. L7:1287 weakly suppresses
  20. L7:1198 weakly suppresses
  21. L5:2141 weakly excites
  22. L6:2635 weakly excites
  23. L6:402 weakly excites
  24. L5:305 weakly suppresses

Feeds later layers

  1. L9:143 suppresses
  2. L9:139 excites
  3. L11:58 suppresses web address colon in urls
  4. L11:513 suppresses words inside idioms and fixed phrases
  5. L10:42 suppresses
  6. L15:397 suppresses
  7. L18:579 weakly suppresses
  8. L13:1659 weakly suppresses
  9. L9:367 weakly excites
  10. L10:152 weakly suppresses
  11. L13:777 weakly suppresses
  12. L9:270 weakly excites
  13. L19:1774 weakly suppresses
  14. L9:58 weakly excites
  15. L15:1483 weakly excites
  16. L9:326 weakly suppresses
  17. L10:67 weakly suppresses
  18. L14:1129 weakly excites
  19. L10:404 weakly suppresses
  20. L21:2402 weakly suppresses
  21. L13:1575 weakly excites
  22. L10:564 weakly excites
  23. L13:1615 weakly suppresses
  24. L18:322 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 -2.86 and -2.751 values between -1.93 and -1.811 values between -1.81 and -1.692 values between -1.69 and -1.574 values between -1.57 and -1.461 values between -1.46 and -1.344 values between -1.34 and -1.225 values between -1.22 and -1.118 values between -1.11 and -0.994 values between -0.99 and -0.878 values between -0.87 and -0.7510 values between -0.75 and -0.6425 values between -0.64 and -0.5229 values between -0.52 and -0.4083 values between -0.40 and -0.28182 values between -0.28 and -0.17703 values between -0.17 and -0.051825 values between -0.05 and 0.07643 values between 0.07 and 0.19210 values between 0.19 and 0.3085 values between 0.30 and 0.4221 values between 0.42 and 0.5418 values between 0.54 and 0.656 values between 0.65 and 0.775 values between 0.77 and 0.896 values between 0.89 and 1.016 values between 1.01 and 1.122 values between 1.12 and 1.245 values between 1.24 and 1.361 values between 1.36 and 1.471 values between 1.47 and 1.592 values between 1.71 and 1.83offnegativepositiveoften

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. anja
  2. @stop
  3. zac
  4. 经济学家
  5. StringWriter
  6. signings
  7. .EndsWith
  8. (Locale
  9. polls
  10. avadoc

Suppresses

  1. �
  2. �
  3. untary
  4. eing
  5. Carr
  6. claro
  7. ying
  8. RTOS
  9. cessive
  10. Real