Neuron Explorer

Qwen3-0.6B

'is' or 'are' after 'what X'

A neuron in layer 18 of Qwen3-0.6B. Its strongest snippets are fairly consistent: several of them peak on the same token. It is featured, so it has steering outputs.

What this neuron does

This neuron lights up on the verb 'is' or 'are' in phrases like 'what a bond is' or 'what cryptocurrencies are', where someone is about to get a definition. Turning it up did not produce definitions; the model's text drifted into repeated filler words.

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

    ⏎Before we dive into calculating the current bond price with our bond valuation calculator, let's take some time to talk about what a bond is. A bond is one of the most prevalent fixed-income securities. When an entity

  2. #2 peak

    ation to a Child⏎When a deceased family member or friend is cremated or already⏎has been cremated, your child may want to know what cremation is.⏎In answering your child's questions about cremation, keep your⏎explanation

  3. #3 peak

    no way a<|endoftext|>If you’ve been around these few months, chances are good that you at least have a passing knowledge of what cryptocurrencies are. You may even have heard of names like Bitcoin or maybe Ethereum to name a

  4. #4 peak

    a basic picture of the relationship between two variables and are useful when showing a side-by-side comparison of various survey questions. Here we will go over what cross tabs are used for, how to set them up, and what the output will look like

  5. #5 peak

    is destined to play a role in the future of education, others would point out that it’s already here. Here is a lowdown on what Augmented Reality is and some ways it can be used in the classroom. Augmented Reality (AR

  6. #6 peak

    huge losses. But worry not. This calculator is built to make the process easier for you.⏎We have written this article to help you understand what a bond price is and how to price a bond using the bond price formula. We will also demonstrate

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: if, instead, I, as, and, as, and, as, and, as, and, as, and, and, and, and, and, and, and, and, and, and, and, and,

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

    the expressions of V(t), I(t), the trigonometric formula sin(X)sin(Y)=1/2(cos(X-Y)-cos(X+Y)), and the fact that (Vmax

  2. #2 peak

    Institute are reflecting on gratitude and what it means for the development of emotional intelligence in young children. As Robert Emmons, Ph.D. the world’s leading scientific expert on gratitude. a professor of psychology and founding editor-in-chief of The Journal of

  3. #3 peak

    of Marine Fisheries, underscored the significance of the striped bass fishery. “It is the recreational fishery in Massachusetts,” he said. “We live and die by the health of striped bass, so we are watching the stocks like hawks

  4. #4 peak

    escape.<|endoftext|>At the end of this lesson, students will be able to:⏎- Identify a geometric sequence.⏎- Graph a geometric sequence.⏎- Solve real-world problems involving geometric sequences.⏎Terms introduced in this lesson:⏎Teaching Strategies and Tips⏎

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. L14:282 suppresses
  2. L7:6 excites
  3. L15:1601 suppresses
  4. L6:101 excites
  5. L16:530 suppresses
  6. L17:2980 suppresses
  7. L16:509 suppresses
  8. L13:1550 suppresses
  9. L6:2503 excites
  10. L6:2477 excites
  11. L17:112 suppresses
  12. L15:286 suppresses
  13. L9:1607 suppresses
  14. L14:177 suppresses
  15. L17:2852 suppresses
  16. L5:53 excites
  17. L9:69 suppresses
  18. L13:210 excites
  19. L12:1198 excites
  20. L10:217 excites
  21. L14:3042 suppresses
  22. L14:713 suppresses
  23. L13:2983 suppresses
  24. L17:354 weakly suppresses

Feeds later layers

  1. L20:353 strongly excites
  2. L19:211 strongly suppresses
  3. L19:608 strongly excites
  4. L20:202 excites
  5. L20:328 excites
  6. L19:1201 excites
  7. L20:756 excites
  8. L20:79 suppresses
  9. L21:243 excites
  10. L20:132 excites
  11. L20:34 suppresses
  12. L19:1496 excites
  13. L21:117 suppresses
  14. L19:83 excites
  15. L20:625 suppresses
  16. L19:223 suppresses
  17. L24:78 suppresses
  18. L19:195 suppresses
  19. L19:1468 excites
  20. L25:334 excites
  21. L20:16 suppresses
  22. L25:210 suppresses
  23. L19:2100 excites
  24. L20:2552 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.

1 values between -1.90 and -1.761 values between -1.76 and -1.621 values between -1.62 and -1.473 values between -1.47 and -1.331 values between -1.33 and -1.192 values between -1.19 and -1.054 values between -1.05 and -0.9013 values between -0.90 and -0.7641 values between -0.76 and -0.62112 values between -0.62 and -0.48292 values between -0.48 and -0.33556 values between -0.33 and -0.19932 values between -0.19 and -0.051097 values between -0.05 and 0.09467 values between 0.09 and 0.24173 values between 0.24 and 0.3870 values between 0.38 and 0.5234 values between 0.52 and 0.6625 values between 0.66 and 0.8118 values between 0.81 and 0.9513 values between 0.95 and 1.0914 values between 1.09 and 1.233 values between 1.23 and 1.388 values between 1.38 and 1.525 values between 1.52 and 1.666 values between 1.66 and 1.803 values between 1.80 and 1.953 values between 2.09 and 2.231 values between 2.23 and 2.372 values between 2.37 and 2.523 values between 2.52 and 2.661 values between 2.80 and 2.941 values between 3.09 and 3.231 values between 3.23 and 3.371 values between 3.65 and 3.80offnegativepositiveoften

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. ,
  3. .
  4. ;
  5. }
  6. :
  7. #
  8. =
  9. aml
  10. ،

Suppresses

  1. equipments
  2. ..."); ⏎
  3. ettel
  4. aphrag
  5. iever
  6. ...)⏎
  7. letal
  8. apping
  9. ichever
  10. ᾷ