Neuron Explorer

Qwen3-0.6B

equals sign or 'is' in formulas

A neuron in layer 14 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 fires on the equals sign or the word 'is' in maths and code, right after a variable name, as in 'vx = ' or 'Step angle ='. Turning it up pushed the writing toward odd number-heavy or unrelated text.

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

    = vsin(theta) Immediately after take off : vx = 2cos25 = 1.8126 vy =

  2. #2 peak

    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, where c = (10/9)2 ≈

  3. #3 peak

    angles of 30, 15, 5, 2.5, 2, and 1.8° per step.⏎Step angle = 360°/No of steps per revolution⏎Where m = Number

  4. #4 peak

    DataTableCollection.⏎Private Sub GetTableByName() ' Presuming a DataGrid is displaying more than one table, ' get its DataSet. Dim thisDataSet As DataSet = CType(DataGrid1.DataSource, DataSet) ' Get the DataTableCollection. Dim tables

  5. #5 peak

    I’ll keep it that way. That just means that I have to make sure that I use gravity as a negative number, since gravity always acts down.⏎vf = vi + at⏎= 56.3m/s + (-9.

  6. #6 peak

    function⏎unorm = sqrt(u(1)*u(1)+u(2)*u(2)); %calculate intermediate value⏎udot(1) = -u(1)/R + u(2) - un<|endoftext|>Thanks for

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 idea of a modern-day "evidence for the universe that we can say it's right at the core of our universe when the universe is at the end of the universe when the universe is at the time of our existence" was a

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

    Grid is displaying more than one table, ' get its DataSet. Dim thisDataSet As DataSet = CType(DataGrid1.DataSource, DataSet) ' Get the DataTableCollection. Dim tablesCollection As DataTableCollection = thisDataSet.Tables ' Get a specific table by

  2. #2 peak

    ⏎t0=0; %set the initial time to zero⏎tend=1; %set the final time to 100⏎

  3. #3 peak

    size) to determine the next point. For example, if the gradient magnitude is 2.5 and the learning rate is 0.01, then the gradient descent algorithm will pick the next point 0.025 away from

  4. #4 peak

    + 314⏎Are the denominators equivalent? Yes, they are.⏎Now we can add the cardinal numbers and the numerators. First the cardinal numbers. That is 2 + 3 = 5. And then

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. L13:1147 suppresses
  2. L13:1988 excites
  3. L12:1865 suppresses
  4. L12:1973 weakly excites
  5. L12:2101 weakly suppresses
  6. L13:1591 weakly suppresses
  7. L12:1254 weakly suppresses
  8. L10:1924 weakly suppresses
  9. L13:2345 weakly excites
  10. L12:2781 weakly suppresses
  11. L6:12 weakly excites
  12. L13:1741 weakly excites
  13. L13:2954 weakly excites
  14. L12:737 weakly suppresses
  15. L13:1561 weakly excites
  16. L10:1625 weakly suppresses
  17. L13:1137 weakly suppresses
  18. L2:808 weakly excites
  19. L13:2424 weakly excites
  20. L5:1566 weakly suppresses
  21. L13:78 weakly excites
  22. L8:74 weakly suppresses
  23. L13:960 weakly excites
  24. L12:1175 weakly excites

Feeds later layers

  1. L15:2583 suppresses
  2. L15:1854 suppresses
  3. L16:1514 suppresses
  4. L16:2786 excites
  5. L16:789 excites
  6. L15:1642 excites
  7. L16:1995 excites
  8. L16:1207 excites
  9. L16:1219 excites
  10. L15:91 excites
  11. L15:1977 suppresses
  12. L17:3024 suppresses
  13. L15:2432 weakly suppresses
  14. L15:1795 weakly excites
  15. L18:2631 weakly suppresses
  16. L19:2068 weakly excites
  17. L15:2554 weakly suppresses
  18. L15:2211 weakly excites
  19. L15:1702 weakly excites
  20. L18:1537 weakly suppresses
  21. L18:115 weakly excites
  22. L16:616 weakly suppresses
  23. L16:702 weakly excites
  24. L16:2971 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.

3 values between -0.77 and -0.721 values between -0.72 and -0.673 values between -0.67 and -0.638 values between -0.63 and -0.587 values between -0.58 and -0.5322 values between -0.53 and -0.4825 values between -0.48 and -0.4351 values between -0.43 and -0.3881 values between -0.38 and -0.34118 values between -0.34 and -0.29137 values between -0.29 and -0.24272 values between -0.24 and -0.19306 values between -0.19 and -0.14433 values between -0.14 and -0.09429 values between -0.09 and -0.05529 values between -0.05 and 0.00453 values between 0.00 and 0.05355 values between 0.05 and 0.10258 values between 0.10 and 0.15165 values between 0.15 and 0.20113 values between 0.20 and 0.2460 values between 0.24 and 0.2933 values between 0.29 and 0.3419 values between 0.34 and 0.3912 values between 0.39 and 0.449 values between 0.44 and 0.492 values between 0.49 and 0.531 values between 0.53 and 0.581 values between 0.63 and 0.681 values between 0.87 and 0.921 values between 1.11 and 1.16offnegativepositiveoften

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. ESH
  2. amiento
  3. uble
  4. uario
  5. стрел
  6. изации
  7. _follow
  8. entifier
  9. ahan
  10. phép

Suppresses

  1. total
  2. 总
  3. totaled
  4. .total
  5. 总投资
  6. curring
  7. 保持
  8. ,total
  9. total
  10. 总体