Neuron Explorer

Qwen3-0.6B

unclear: small words in maths and how-to text

A neuron in layer 15 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

The positive side is weak and fires on small words like 'a', 'the' and 'to' in maths and instructions, such as 'percentage of a whole number'. The much stronger side is negative and fires on words like 'starts' in graph definitions. Turning it up gave unrelated, repetitive text.

Written by Claude from this neuron's activations and steering outputs. Confidence: low. Think it is wrong? Flag it (opens an email to the maker with the neuron already filled in).

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

    Jersey Turnpike. He was 86.⏎There are several different ways to solve problems involving percentages, decimals, and fractions. Watch as find the percentage of a whole number.<|endoftext|>1. How does the U.S. manage the environmental legacy

  2. #2 peak

    , it can also be modified manually in the top menu.⏎Now, the formula to get the percentages is basically: the letter and the cell number of the general amount, divided by the letter and the

  3. #3 peak

    Strategies and Tips⏎Use this lesson to show how exponential functions and geometric sequences are related.⏎- In a geometric sequence, terms are found by multiplying the same constant to the previous term.⏎- For exponential functions, when the input variable is increased by

  4. #4 peak

    function can be used to determine, based on the employee's name, the position in the array, or table. Check the array in that position and see if the record is there.⏎Hash functions have a big science behind it because as you may

  5. #5 peak

    make use of a mark percentage calculator to calculate your marks. If you’re unsure of how to write a number, try a calculator.⏎To find the percentage of a number, you should first know how many decimal places it falls under. You can

  6. #6 peak

    to each other (parallel to the axis of symmetry as well).⏎: A parabola is the set of all points in a plane that are equidistant from the focus and the directrix of the parabola. [Adopted

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 been a participant in a whole (i.e., a total) change in the sense that I have gone from the beginning, so I have gone to the end, so the change in total is zero. So, for example,

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

    a large role in our graph algorithms. A path is some sequence of edges that allows you to travel from one node to another. A cycle is a path that starts and ends at the same node. Sometimes cycles are useful, but they often represent

  2. #2 peak

    flow are what keep a business going. What good are record revenues if you’re spending more money than you’re taking in?⏎Sidebar – My buddy Steve Chou owns a successful online wedding

  3. #3 peak

    . How many new shapes can you make?⏎Can you shunt the trucks so that the Cattle truck and the Sheep⏎truck change places and the Engine is back on the main line?⏎In how many ways can you fit two of these yellow

  4. #4 peak

    negative, which sum to zero.⏎The positive cycle may move charge carriers, such as electrons. But then the negative cycle pulls the charges back to where they started. To reliably control the quantum information, an asymmetric light wave is needed.⏎“The optimum

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. L12:2854 excites
  2. L9:1623 suppresses
  3. L6:2090 suppresses
  4. L13:2770 suppresses
  5. L13:2309 excites
  6. L6:1145 excites
  7. L13:990 suppresses
  8. L12:2541 excites
  9. L11:2410 excites
  10. L14:2445 excites
  11. L8:1526 excites
  12. L13:2816 weakly suppresses
  13. L9:2391 weakly excites
  14. L7:2309 weakly excites
  15. L13:2942 weakly excites
  16. L10:141 weakly excites
  17. L7:536 weakly suppresses
  18. L7:245 weakly excites 'even' in 'and even' lists
  19. L11:951 weakly suppresses
  20. L12:2192 weakly suppresses
  21. L11:557 weakly excites
  22. L7:513 weakly suppresses
  23. L11:543 weakly excites
  24. L6:2517 weakly excites

Feeds later layers

  1. L16:2990 excites
  2. L16:2123 suppresses
  3. L16:2756 suppresses
  4. L19:746 suppresses
  5. L18:2955 weakly suppresses
  6. L16:1376 weakly excites
  7. L21:2233 weakly suppresses
  8. L18:969 weakly suppresses
  9. L18:1710 weakly suppresses
  10. L17:1726 weakly suppresses
  11. L17:2351 weakly suppresses
  12. L16:2894 weakly suppresses
  13. L20:1768 weakly suppresses
  14. L22:2541 weakly suppresses
  15. L17:2887 weakly excites
  16. L20:1362 weakly suppresses
  17. L22:2920 weakly excites
  18. L23:2539 weakly suppresses
  19. L18:2910 weakly suppresses
  20. L18:2643 weakly suppresses
  21. L20:2576 weakly suppresses
  22. L23:1897 weakly suppresses
  23. L18:2879 weakly excites
  24. L16:1279 weakly 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 -3.86 and -3.741 values between -3.25 and -3.131 values between -3.01 and -2.881 values between -2.03 and -1.901 values between -1.90 and -1.781 values between -1.41 and -1.291 values between -1.29 and -1.174 values between -1.17 and -1.054 values between -1.05 and -0.925 values between -0.92 and -0.805 values between -0.80 and -0.684 values between -0.68 and -0.5610 values between -0.56 and -0.4327 values between -0.43 and -0.3187 values between -0.31 and -0.19311 values between -0.19 and -0.07995 values between -0.07 and 0.061065 values between 0.06 and 0.18819 values between 0.18 and 0.30401 values between 0.30 and 0.42126 values between 0.42 and 0.5534 values between 0.55 and 0.673 values between 0.67 and 0.791 values between 0.91 and 1.04offnegativepositiveoften

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. 分辨率
  2. PLEX
  3. ⌠
  4. urgently
  5. 部分内容
  6. ropdown
  7. 紧迫
  8. 慑
  9. 霸道
  10. urgent

Suppresses

  1. cancelled
  2. vanish
  3. canceled
  4. ancellation
  5. 消失
  6. neutral
  7. canc
  8. vanished
  9. cancel
  10. Cancellation