Neuron Explorer

Qwen3-0.6B

the '3' in '3D' and '2D'

A neuron in layer 23 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 number before a 'D', as in '3D' or '2-D' for three-dimensional things. When turned up, the model's replies filled with a repeated capital D, so it is tied to the letter D after a number.

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

    separate GPU is less important than with more complicated or moving images.⏎In video games, a separate graphics card helps to turn a set of instructions into a complex 3D image with surface textures and shading that are constantly changing. The 3D

  2. #2 peak

    and some ways it can be used in the classroom. Augmented Reality (AR) is the combination of reality and virtual reality, and allows you to put 3D content and images into the real world. AR has many applications in various fields

  3. #3 peak

    laser scanning.⏎In the overall Scan-to-BIM process, a laser scanner is used to acquire a precise 3D scan of the project’s real-world conditions.⏎The scan data is then loaded

  4. #4 peak

    IM creates a data-rich, linked, and constructible workflow. Having precise measurements to work with is essential for creating accurate models. The accuracy of measurements and 3D model content in BIM projects has reached new heights thanks to 3D

  5. #5 peak

    our right when it starts and know its next move, maybe we can also stop it from growing at all.⏎EMBL scientists have created a new, realistic 3D testbed that could help them achieve that goal by studying cancer cells as they

  6. #6 peak

    US decarbonization target. The Ministry of Energy’s solar energy research shows that the new installed capacity in the next five years<|endoftext|>Holograms help build 3D nanostructures⏎Mar 1, 2002⏎A new

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 use of232222222222222222222222222222222222222222222

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

    takes place in the mid-18th century, Rzewuski created a gallery of types from the Polish gentry—bullies, drunkards, squabblers, and ignoramuses from Prince Radziwiłł’s ret

  2. #2 peak

    extreme makeover. Back in the 1970s, squatters, hawkers and manufacturing industries crowded the banks of the river, leading to

  3. #3 peak

    about so they can see the animals in their native habitats, rather than simply pictures in a textbook.⏎The workshops, which are funded by the Bristol-Myers Squibb Center, are offered free to the teachers. "The goal of the

  4. #4 peak

    , and established farmers and artisans known as Grensboere, or Border Farmers. Amongst the Voortrekkers were poor men too belonging to the squatter or bywoner class. Together these groups were later called Voortre

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. L17:128 excites
  2. L22:747 excites
  3. L22:332 suppresses
  4. L19:723 excites
  5. L15:12 excites
  6. L4:205 weakly excites the digit 2 inside citations
  7. L19:258 weakly suppresses
  8. L20:225 weakly suppresses
  9. L5:1101 weakly excites
  10. L7:229 weakly excites first digit of a split number
  11. L18:1490 weakly suppresses
  12. L0:2289 weakly suppresses
  13. L8:463 weakly excites
  14. L7:556 weakly excites
  15. L18:569 weakly excites
  16. L12:713 weakly suppresses
  17. L8:20 weakly suppresses
  18. L13:33 weakly excites
  19. L2:49 weakly suppresses
  20. L20:1823 weakly suppresses
  21. L13:545 weakly suppresses 'of' or 'to' linking two things
  22. L7:303 weakly suppresses
  23. L16:1220 weakly excites
  24. L19:2307 weakly excites

Feeds later layers

  1. L27:552 strongly excites
  2. L26:2996 excites
  3. L27:343 excites
  4. L26:2168 excites
  5. L26:901 excites
  6. L25:3008 excites
  7. L25:224 excites
  8. L25:279 excites
  9. L25:617 excites
  10. L24:371 excites
  11. L25:313 excites
  12. L25:543 weakly suppresses
  13. L27:294 weakly excites
  14. L27:1202 weakly excites
  15. L26:1217 weakly excites
  16. L27:1680 weakly excites
  17. L27:1440 weakly excites
  18. L25:379 weakly suppresses
  19. L25:2657 weakly excites
  20. L25:232 weakly excites
  21. L27:1023 weakly excites
  22. L25:1244 weakly suppresses
  23. L25:138 weakly excites
  24. L26:834 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 -22.58 and -21.341 values between -8.98 and -7.752 values between -7.75 and -6.511 values between -6.51 and -5.281 values between -5.28 and -4.041 values between -4.04 and -2.805 values between -2.80 and -1.57105 values between -1.57 and -0.333731 values between -0.33 and 0.9123 values between 0.91 and 2.1412 values between 2.14 and 3.388 values between 3.38 and 4.613 values between 4.61 and 5.852 values between 5.85 and 7.082 values between 7.08 and 8.321 values between 8.32 and 9.561 values between 9.56 and 10.791 values between 12.03 and 13.261 values between 13.26 and 14.502 values between 14.50 and 15.742 values between 19.44 and 20.681 values between 23.15 and 24.391 values between 25.62 and 26.86offnegativepositiveoften

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. D
  2. D
  3. -D
  4. d
  5. Ds
  6. D
  7. 𝐷
  8. *D
  9. [D
  10. ,D

Suppresses

  1. lops
  2. zon
  3. 2
  4. lop
  5. 二十四
  6. 3
  7. tps
  8. neh
  9. 1
  10. apons