Neuron Explorer

Qwen3-0.6B

opening parenthesis in formulas and math

A neuron in layer 11 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 reacts to an opening bracket, most often in maths or formula text such as a calculation or a formula being written out. Turning it up made the writing wander into number-heavy and maths-like text, and turning it down made it repeat itself.

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

    industry are able to sell,” the report said.<|endoftext|>Determining if a Product is Positive or Negative⏎Recall the following properties of multiplication of signed numbers:⏎(positive)(positive) = positive⏎(positive)(negative) = negative⏎(n

  2. #2 peak

    4,000-lb cart with 4 wheels required to carry a 6,000-lb load would use the following calculation:⏎[(4,000 lbs + 6,000 lbs) ×

  3. #3 peak

    providers of capital, or they will invest elsewhere.⏎The following is the WACC calculation formula:⏎WACC = E/V × Re + D/V × Rd × (1 - Tc)⏎Re = cost of equity⏎Rd = cost of

  4. #4 peak

    (t), I(t), the trigonometric formula sin(X)sin(Y)=1/2(cos(X-Y)-cos(X+Y)), and the fact that (Vmax×Imax)/2=Vrms×Irms, it comes

  5. #5 peak

    equal to L. Three Special Limits 1. sin lim 1 x x x = 2. 1 cos lim x x x -= 3. ( ) 1 lim 1 x x x e + = Strategy for Finding Limits

  6. #6 peak

    requirements:⏎3-4 pages in length (does not include title and reference page)⏎1-2 outside sources⏎Why Work with CheggAnswers⏎At CheggAnswers.com, we boast of our quick turnaround that is facilitated by our experienced team

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 21st century, - The 22nd century, - The 23rd century, - The 24th century, - The 25th century, - The

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

    hydrogen fuel cell car fleet may hit top gear within five years, but only if there is enough investment in filling stations, the UK Hydrogen and Fuel Cells Association (UK HFCA) told Reuters on Friday.⏎Fuel cells convert hydrogen into electricity,

  2. #2 peak

    of Alaskan Matthew Wayne Henry for possessing a homemade machine gun Thursday, the Ninth Circuit held:⏎We agree with the reasoning of our sister circuits that machine guns are “dangerous and unusual weapons” that are not protected by the Second Amendment. An

  3. #3 peak

    Mina MJ, Leng Y, Elledge SJ, Niebuhr DW, Scher AI, Munger KL and Ascherio A⏎Multiple sclerosis (MS) is a chronic inflammatory demyelinating disease of the central nervous system

  4. #4 peak

    the<|endoftext|>Green Infrastructure for Clean Water Act⏎January 22 2010 | Know Your H20,⏎The Green Infrastructure for Clean Water Act (HR

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. L10:2562 excites
  2. L10:2993 excites
  3. L10:1883 excites closing parenthesis after a number or unit
  4. L10:2104 excites opening parenthesis in formulas and citations
  5. L9:1108 weakly suppresses
  6. L2:2341 weakly excites
  7. L9:2099 weakly suppresses
  8. L8:189 weakly suppresses
  9. L5:720 weakly suppresses
  10. L0:479 weakly suppresses
  11. L3:2123 weakly excites
  12. L2:1219 weakly suppresses
  13. L0:2961 weakly suppresses
  14. L6:144 weakly excites
  15. L8:2143 weakly excites
  16. L9:838 weakly excites
  17. L8:307 weakly suppresses
  18. L3:2974 weakly suppresses
  19. L8:2308 weakly excites
  20. L5:2177 weakly suppresses
  21. L5:708 weakly excites
  22. L5:436 weakly suppresses
  23. L9:270 weakly suppresses
  24. L6:2783 weakly excites

Feeds later layers

  1. L13:1698 weakly excites
  2. L15:1363 weakly suppresses
  3. L12:1278 weakly excites
  4. L15:2012 weakly suppresses
  5. L13:2444 weakly excites
  6. L16:271 weakly excites
  7. L17:113 weakly excites
  8. L13:2903 weakly excites
  9. L14:1274 weakly excites
  10. L13:1985 weakly excites
  11. L18:2477 weakly excites
  12. L14:646 weakly excites
  13. L12:1287 weakly excites
  14. L24:2946 weakly excites
  15. L16:995 weakly excites
  16. L12:2252 weakly suppresses
  17. L12:2666 weakly excites
  18. L17:1595 weakly excites
  19. L27:2997 weakly excites
  20. L13:2065 weakly suppresses
  21. L20:609 weakly excites
  22. L12:473 weakly excites
  23. L20:2676 weakly suppresses
  24. L15:1794 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 -2.61 and -2.531 values between -1.57 and -1.491 values between -1.25 and -1.171 values between -0.93 and -0.851 values between -0.77 and -0.693 values between -0.69 and -0.613 values between -0.61 and -0.535 values between -0.53 and -0.459 values between -0.45 and -0.3725 values between -0.37 and -0.2995 values between -0.29 and -0.21252 values between -0.21 and -0.12471 values between -0.12 and -0.041082 values between -0.04 and 0.04855 values between 0.04 and 0.12591 values between 0.12 and 0.20314 values between 0.20 and 0.28134 values between 0.28 and 0.3650 values between 0.36 and 0.4410 values between 0.44 and 0.524 values between 0.52 and 0.60offnegativepositiveoften

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. ilated
  2. annels
  3. ionario
  4. rometer
  5. _cube
  6. pooled
  7. ilyn
  8. averse
  9. angrily
  10. häuser

Suppresses

  1. ality
  2. “
  3. 这些问题
  4. [,
  5. [
  6. ока
  7. []
  8. [,]
  9. dir
  10. (s