Neuron Explorer

Qwen3-0.6B

'to' before a verb of purpose

A neuron in layer 22 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 lights up on the word 'to' when it comes before an action, as in 'how to calculate' or 'used to find'. It is predicting that a verb like 'find' or 'calculate' comes next. When it was turned up, the model got stuck repeating 'find' and 'a good way to find'.

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

    should be called getVolumeOfSphere(), take one parameter called radius and apply the relevant formula to calculate and return the volume of a sphere based on the given radius.⏎- Your

  2. #2 peak

    (9/10)k-2, where k is the number of digits of 2n. The number of digits can be estimated as n log102. Thus, the probability is about cxn

  3. #3 peak

    do the rest, including converting the decimal to a percentage.⏎To find the percentage value of a mark, you’ll need to know its total mark. In order to find this, use basic arithmetic. Divide the mark by the highest possible mark and

  4. #4 peak

    .net, a new website, provides goal writing guidance, especially useful for early intervention and children on the autistic spectrum disorder.<|endoftext|>The online WACC Calculator is used to calculate the weighted average cost of capital (WACC).⏎In finance, The weighted

  5. #5 peak

    a deep breath and let us take it one step at a time. Click and begin.⏎GCF and LCM by Prime Factorization⏎A straight forward way to find the greatest common factor (GCF) and least common multiple (LCM

  6. #6 peak

    important, if not the most important, metrics in valuing the bond.⏎Now, let's look at some examples to understand how to calculate bond prices.⏎How to calculate the bond price? The bond price formula⏎Bond price is calculated as the

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: I have a 2000 model, 2000 2000, 2000, 2000, 2000, 2000, 20

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 the slope. The discrete model of exponential growth is ; where the first term is , and is the common ratio. The continuous model is .⏎1. Find the common ratio of each geometric sequence.⏎

  2. #2 peak

    5 Y 2 . Solution : X =- 5 W c) Find some solution of A X = 3 Y 1- 5 Y

  3. #3 peak

    It is true that today we have online limit calculators that can help in limit calculation. In this post we are going to discuss the reasons as to why find the limit calculation is a tough task especially for high school students.⏎Teaching math to

  4. #4 peak

    x x = 2. 1 cos lim x x x -= 3. ( ) 1 lim 1 x x x e + = Strategy for Finding Limits: 1. Try direct substitution: One of three things will happen:

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. L21:234 suppresses
  2. L21:2199 suppresses
  3. L16:2597 excites
  4. L4:3053 suppresses
  5. L20:700 excites
  6. L18:1579 excites
  7. L17:2903 excites
  8. L21:1651 weakly excites
  9. L21:317 weakly excites
  10. L1:710 weakly suppresses
  11. L18:2999 weakly excites
  12. L17:2257 weakly suppresses
  13. L1:2206 weakly suppresses
  14. L5:78 weakly excites
  15. L17:789 weakly excites
  16. L15:1445 weakly excites
  17. L15:2437 weakly suppresses opening parenthesis after a noun
  18. L17:1912 weakly excites
  19. L18:2584 weakly suppresses
  20. L18:2628 weakly suppresses
  21. L4:713 weakly excites
  22. L12:1390 weakly excites
  23. L12:354 weakly excites
  24. L13:2110 weakly excites

Feeds later layers

  1. L27:1284 strongly excites
  2. L24:1909 excites
  3. L23:540 excites
  4. L27:1207 excites
  5. L26:2988 excites
  6. L25:660 excites
  7. L24:1302 excites
  8. L24:769 excites
  9. L27:2510 excites
  10. L23:1325 excites
  11. L23:1432 excites
  12. L23:494 excites
  13. L26:1422 weakly excites
  14. L24:837 weakly excites
  15. L25:1335 weakly excites
  16. L24:1068 weakly suppresses
  17. L25:2811 weakly excites
  18. L24:2247 weakly excites
  19. L26:1282 weakly excites
  20. L26:656 weakly suppresses
  21. L23:3066 weakly suppresses
  22. L24:545 weakly excites
  23. L23:136 weakly excites
  24. L25:1035 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.

1 values between -5.34 and -4.991 values between -4.27 and -3.912 values between -3.55 and -3.193 values between -2.83 and -2.473 values between -2.47 and -2.114 values between -2.11 and -1.757 values between -1.75 and -1.4023 values between -1.40 and -1.0480 values between -1.04 and -0.68293 values between -0.68 and -0.321230 values between -0.32 and 0.041463 values between 0.04 and 0.40626 values between 0.40 and 0.76133 values between 0.76 and 1.1215 values between 1.12 and 1.485 values between 1.48 and 1.843 values between 1.84 and 2.204 values between 2.20 and 2.552 values between 2.55 and 2.912 values between 2.91 and 3.274 values between 3.63 and 3.991 values between 5.07 and 5.431 values between 6.14 and 6.501 values between 6.50 and 6.861 values between 8.66 and 9.02offnegativepositiveoften

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. finding
  2. 计算
  3. find
  4. find
  5. computing
  6. Find
  7. finding
  8. 求
  9. finds
  10. Find

Suppresses

  1. pered
  2. AFX
  3. brush
  4. 听取
  5. izzly
  6. 回头看
  7. 诃
  8. abyte
  9. serialVersionUID
  10. مل