Neuron Explorer

Qwen3-0.6B

"of" after "components" or "causes"

A neuron in layer 1 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 reacts to the word "of" when it follows a noun like "components", "causes" or "drivers", as in "the main components of". Turning it up did not produce a clear change in the model's writing.

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

    main components of the website are Understanding Science 101, a “primer on the nature

  2. #2 peak

    in the journal Nature Physics.<|endoftext|>Yoga is a form of exercise that focuses on strength, flexibility and breathing to boost physical and mental wellbeing. The main components of a yoga class are asanas (postures) and movements designed to increase strength

  3. #3 peak

    of Geospatial data, Linked Open data and other environmental data (maps) for educational and gaming purposes. The main components of the environment introduced are – water, air, soil, forests, nature protection,

  4. #4 peak

    able and explore its significance in the realm of finance.⏎- 1. Understanding Bilan Comptable: A Comprehensive Guide⏎- 2. Key Components of a Bilan Comptable: Assets, Liabilities, and Equity⏎-

  5. #5 peak

    . Moreover, the symptoms of this syndrome might include puffiness around eyes, ankles, and feet, weight gain, and frothy urine.⏎What are the causes of Kidney Failure?⏎- Kidney stones⏎- Polycystic kidney disease

  6. #6 peak

    innate in humans to feel empathy toward others, and it happens mostly involuntarily. However, some experiences trigger empathy more accurately and intensely.⏎What Are the Components of Empathy

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: A new beginning, the beginning of a new day, the beginning of a new day. So, the user has provided text about, and the user has been thinking about. Let me know I am thinking about, and the user has been

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

    , and Threats⏎- My Strategic Plan: Internal and External Analysis⏎- University of Missouri Extension: Marketing⏎- Advanced Performance Institute: What is a Key Performance Indicator (KPI)?⏎- Virtual Advisor, Inc.: Analyzing Your Competition⏎

  2. #2 peak

    it continues to provide a key source of data, context and actionable insights for the U.S. cotton industry.⏎The peer-reviewed NIR analyzes sustainability trends across five key performance indicators (KPIs) for U.S. commodity crops, including cotton.

  3. #3 peak

    countries in the world for direct foreign industrial investment. However, UK manufacturing firms and industries often significantly lag behind their overseas competitors in terms of productivity and various other key performance measures. Of late, such things

  4. #4 peak

    So if you'd like to learn something about the best strategy to use when teaching your students to fly, why not just ask them to tell you?⏎The next time you're introduced to a student, spend some time chatting with him about some of

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. L0:1609 weakly suppresses
  2. L0:2887 weakly excites
  3. L0:178 weakly suppresses
  4. L0:1248 weakly excites
  5. L0:2093 weakly excites
  6. L0:1224 weakly suppresses
  7. L0:778 weakly excites
  8. L0:2920 weakly suppresses
  9. L0:2964 weakly excites
  10. L0:512 weakly suppresses
  11. L0:645 weakly excites
  12. L0:186 weakly excites
  13. L0:2198 weakly suppresses
  14. L0:1870 weakly suppresses
  15. L0:2660 weakly suppresses
  16. L0:1289 weakly suppresses
  17. L0:244 weakly suppresses
  18. L0:252 weakly excites
  19. L0:2371 weakly excites
  20. L0:1301 weakly suppresses
  21. L0:1493 weakly excites
  22. L0:35 weakly suppresses
  23. L0:953 weakly suppresses
  24. L0:2074 weakly suppresses

Feeds later layers

  1. L4:698 weakly excites
  2. L17:2311 weakly suppresses
  3. L14:890 weakly excites
  4. L13:2876 weakly excites
  5. L13:60 weakly suppresses
  6. L8:2711 weakly suppresses
  7. L12:1400 weakly excites
  8. L17:978 weakly suppresses
  9. L17:1685 weakly suppresses
  10. L19:603 weakly excites
  11. L19:900 weakly suppresses opening parenthesis after a term
  12. L7:297 weakly excites
  13. L21:833 weakly excites
  14. L11:1592 weakly excites
  15. L2:330 weakly excites
  16. L3:737 weakly excites
  17. L4:2401 weakly suppresses
  18. L12:1870 weakly excites
  19. L5:2546 weakly excites
  20. L25:646 weakly suppresses measurement words before a value
  21. L20:1525 weakly suppresses
  22. L14:1638 weakly suppresses
  23. L7:2432 weakly excites
  24. L12:4 weakly suppresses document-end markers, mixed

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.

4 values between -0.62 and -0.553 values between -0.55 and -0.474 values between -0.47 and -0.4014 values between -0.40 and -0.3231 values between -0.32 and -0.25106 values between -0.25 and -0.17356 values between -0.17 and -0.09967 values between -0.09 and -0.021566 values between -0.02 and 0.06525 values between 0.06 and 0.13170 values between 0.13 and 0.2163 values between 0.21 and 0.2825 values between 0.28 and 0.3616 values between 0.36 and 0.4310 values between 0.43 and 0.515 values between 0.51 and 0.588 values between 0.58 and 0.668 values between 0.66 and 0.734 values between 0.73 and 0.812 values between 0.81 and 0.883 values between 0.88 and 0.962 values between 0.96 and 1.031 values between 1.03 and 1.112 values between 1.11 and 1.181 values between 1.18 and 1.263 values between 1.26 and 1.331 values between 1.33 and 1.413 values between 1.41 and 1.491 values between 1.49 and 1.561 values between 1.71 and 1.791 values between 2.01 and 2.092 values between 2.31 and 2.39offnegativepositiveoften

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. новости
  2. ской
  3. sse
  4. aida
  5. Cause
  6. кла
  7. 这款车
  8. рег
  9. lagen
  10. TagName

Suppresses

  1. tempt
  2. imum
  3. intervals
  4. rompt
  5. conven
  6. *"
  7. 貌
  8. inen
  9. spontaneous
  10. IMAL