Neuron Explorer

Qwen3-0.6B

list-style essay phrasing, weak

A neuron in layer 10 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 weakly responds to the middle of long, polished explanatory sentences, especially where a comma, connective or next word is about to appear in an essay about work or skills. The signal is small, and turning it up did not give 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

    perfect when it comes to reading circuit diagrams. So don't be afraid to get your hands dirty and start tinkering! With a little bit of patience and dedication, you'll soon start to recognize the patterns and symbols and become an expert in no

  2. #2 peak

    reusable code, efficient algorithms, and proper documentation enhances the app’s maintainability and scalability. Embracing agile development methodologies and collaboration tools promotes effective teamwork and streamlines the development process.⏎Furthermore, continuous testing, feedback loops, and iterative improvements ensure

  3. #3 peak

    old problems.⏎However, there are also disadvantages to smart work. First, it can lead to a lack of focus on the details, as individuals may prioritize efficiency over accuracy.

  4. #4 peak

    working tirelessly towards the desired outcome, often without taking breaks or shortcuts. This approach is often seen as noble and honorable, as it requires dedication, determination, and a strong work ethic.⏎One of the main advantages of hard work is that it can

  5. #5 peak

    equity. By comparing these three components, it is possible to determine the overall net worth of the business. This information is vital for decision-making, financial analysis, and evaluating the company's ability to meet its financial obligations.⏎Assets represent what the company

  6. #6 peak

    a renowned American poet, has captivated readers with her profound exploration of nature and the human experience. Her poems often evoke a sense of wonder and contemplation, inviting readers to reflect on their own lives. In this argumentative essay, we will

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 concept of a mathematical functions related with polynomials to different degrees of 2 and 4 and 5 and 5 and 5. For example, the 5th order terms for polynomials related with functions like x^2

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

    functional applications. It provides the means to create engaging user interfaces, implement complex algorithms, and integrate various features seamlessly. By leveraging programming languages such as Java, Swift, or Kotlin, developers can build apps for different platforms like iOS, Android, or

  2. #2 peak

    . They face the challenge of optimizing app performance, ensuring compatibility across multiple devices, and addressing varying screen sizes and resolutions. Additionally, staying up-to-date with emerging technologies and frameworks is crucial to deliver cutting-edge apps that cater to users’ evolving expectations

  3. #3 peak

    ’s quality and usability. Regular updates and bug fixes based on user feedback contribute to long-term success. Lastly, keeping an eye on industry trends, exploring new technologies, and participating in developer communities foster growth and innovation

  4. #4 peak

    muscle group is necessary for optimal muscle growth. For instance, Fonseca et al. (2014) showed that combining different lower-body exercises resulted in muscle growth in the whole quadriceps muscle

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. L9:23 suppresses
  2. L8:642 suppresses
  3. L8:1104 suppresses
  4. L9:218 suppresses
  5. L8:609 suppresses
  6. L8:663 suppresses
  7. L3:157 excites
  8. L3:17 suppresses
  9. L8:2083 suppresses
  10. L8:1415 suppresses
  11. L4:50 excites
  12. L7:276 suppresses
  13. L8:967 excites
  14. L4:121 weakly excites
  15. L9:614 weakly suppresses
  16. L9:302 weakly excites
  17. L7:838 weakly suppresses
  18. L3:177 weakly suppresses
  19. L8:2023 weakly excites
  20. L2:1423 weakly excites
  21. L9:949 weakly excites
  22. L5:475 weakly excites
  23. L9:1573 weakly suppresses
  24. L8:789 weakly suppresses

Feeds later layers

  1. L16:704 strongly suppresses
  2. L14:92 strongly excites
  3. L12:1788 strongly suppresses
  4. L12:1091 strongly suppresses
  5. L12:1005 strongly suppresses
  6. L11:638 strongly suppresses
  7. L17:349 strongly suppresses
  8. L18:232 strongly excites
  9. L11:2500 strongly suppresses
  10. L14:3017 strongly suppresses
  11. L11:1970 suppresses
  12. L17:2272 suppresses
  13. L14:1266 suppresses
  14. L14:2049 suppresses
  15. L12:1779 suppresses weak end of document signal
  16. L15:67 excites
  17. L12:1696 suppresses
  18. L13:40 excites
  19. L17:681 suppresses
  20. L12:1957 suppresses
  21. L17:516 suppresses
  22. L15:634 suppresses
  23. L22:256 suppresses
  24. L12:2684 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 -0.81 and -0.771 values between -0.74 and -0.703 values between -0.59 and -0.552 values between -0.51 and -0.471 values between -0.44 and -0.401 values between -0.40 and -0.362 values between -0.36 and -0.335 values between -0.33 and -0.295 values between -0.29 and -0.257 values between -0.25 and -0.2110 values between -0.21 and -0.1819 values between -0.18 and -0.1436 values between -0.14 and -0.1067 values between -0.10 and -0.06112 values between -0.06 and -0.03297 values between -0.03 and 0.01348 values between 0.01 and 0.05372 values between 0.05 and 0.09370 values between 0.09 and 0.12435 values between 0.12 and 0.16402 values between 0.16 and 0.20362 values between 0.20 and 0.24268 values between 0.24 and 0.27244 values between 0.27 and 0.31188 values between 0.31 and 0.35137 values between 0.35 and 0.3892 values between 0.38 and 0.4255 values between 0.42 and 0.4630 values between 0.46 and 0.5025 values between 0.50 and 0.534 values between 0.53 and 0.572 values between 0.57 and 0.613 values between 0.61 and 0.652 values between 0.65 and 0.68offnegativepositiveoften

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. “
  2. 来讲
  3. “[
  4. ’
  5. (“
  6. “
  7. 过来
  8. ุด
  9. ‘s
  10. ễ

Suppresses

  1. ."⏎⏎⏎⏎
  2. ait
  3. rance
  4. ]⏎⏎⏎⏎
  5. ?"⏎⏎⏎⏎
  6. "⏎⏎⏎⏎
  7. edList
  8. ...⏎⏎⏎⏎
  9. ais
  10. ⏎ ⏎ ⏎ ⏎