Neuron Explorer

Qwen3-0.6B

empty spot before a number or "3D"

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 fires on a blank spot in the text that is directly followed by a number, like "3D software", "3D printer" or "360-degree". Turned down, the model wrote lists of numbers, and turned up it wandered into unrelated sentences.

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

    questions in complete sentences and paragraphs based on 7 days, worldwide (use zoom feature to zoom in or out). Your responses

  2. #2 peak

    get to the part where I show you literally how to design a complete house using a popular free and open source 3D software, we must first learn some designing techniques. Using a very easy

  3. #3 peak

    way of making things involved taking lots of parts and screwing or welding them together. Now a product can be designed on a computer and “printed” on a 3D printer, which creates a solid object by building up successive layers of material

  4. #4 peak

    ’s subjects, and the public congregations and usual places of divine worship; be<|endoftext|>New anti-flu drugs could be on the market within years after scientists create 3D map of part of the virus

  5. #5 peak

    —the popular online file repository for 3-D objects—and expect to change the world. “I think one of the

  6. #6 peak

    which all interactions are reduced to defining the determination of the Matter or Economy in a society: social<|endoftext|>Touring the Milky Way’s a blast with this brand new 360-degree interactive panorama. More than 2 million infrared photos taken by

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: 1. 2. 3. 4. 5. 6.

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

    04, 2010 · LEDC flooding case study Bangladesh 1998 1. LEDC Flooding Case Study: Bangladesh 1998 2. Between July-September 1

  2. #2 peak

    can add the cardinal numbers and the numerators. First the cardinal numbers. That's only one 3. This will remain. And then the numerators. 1 + 3 = 4. We then get the answer 34

  3. #3 peak

    1, data 0, rloc 0, and pref 0. The GLib 0 resource in a shared library corresponds exactly to the code 1 resource in an application.⏎Before going any further, you need to make sure

  4. #4 peak

    cart with 4 wheels required to carry a 6,000-lb load would use the following calculation:⏎[(4,000 lbs + 6,000 lbs) × 1.25] ÷

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. L18:106 strongly suppresses
  2. L19:581 suppresses blank slot before a number
  3. L14:191 excites blank spot before a run-on number
  4. L4:53 excites blank slot before a number
  5. L21:173 excites
  6. L21:339 excites spot just before a year
  7. L16:471 suppresses blank spot before a year
  8. L16:9 suppresses
  9. L18:377 excites number right after 'about' or 'less than'
  10. L5:334 suppresses
  11. L7:229 weakly excites first digit of a split number
  12. L6:0 weakly excites
  13. L16:66 weakly excites
  14. L20:344 weakly excites hyphen inside compound words
  15. L8:30 weakly excites blank slot right before a number
  16. L21:189 weakly suppresses document break token mid-sentence
  17. L20:433 weakly excites
  18. L6:1279 weakly suppresses
  19. L15:12 weakly excites
  20. L17:816 weakly suppresses
  21. L18:48 weakly suppresses
  22. L16:55 weakly excites
  23. L18:1534 weakly suppresses space between month and day number
  24. L11:1601 weakly suppresses

Feeds later layers

  1. L27:170 strongly excites unclear: spaces in mixed-language text
  2. L23:77 excites the space right before a number
  3. L26:423 suppresses
  4. L27:448 excites
  5. L27:237 suppresses
  6. L27:113 excites
  7. L26:172 excites
  8. L26:611 suppresses
  9. L25:33 excites spot where a number is due
  10. L27:306 suppresses
  11. L25:95 suppresses
  12. L27:283 excites the spot right before a number
  13. L27:1017 suppresses
  14. L24:675 suppresses
  15. L26:97 suppresses
  16. L25:736 excites
  17. L26:542 suppresses
  18. L27:287 suppresses
  19. L26:2873 suppresses
  20. L26:57 excites
  21. L25:222 suppresses
  22. L25:91 excites
  23. L27:183 suppresses
  24. L27:1652 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 -4.06 and -3.0668 values between -2.05 and -1.053065 values between -1.05 and -0.04740 values between -0.04 and 0.968 values between 0.96 and 1.963 values between 1.96 and 2.972 values between 3.97 and 4.981 values between 4.98 and 5.982 values between 5.98 and 6.981 values between 6.98 and 7.991 values between 8.99 and 10.001 values between 10.00 and 11.001 values between 11.00 and 12.003 values between 12.00 and 13.011 values between 13.01 and 14.012 values between 14.01 and 15.011 values between 15.01 and 16.021 values between 16.02 and 17.021 values between 17.02 and 18.031 values between 19.03 and 20.031 values between 21.04 and 22.041 values between 24.05 and 25.051 values between 33.08 and 34.091 values between 35.09 and 36.09offnegativepositiveoften

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. '/')
  2. ibel
  3. �藏
  4. ionic
  5. 喟
  6. prof
  7. .hxx
  8. ****************************
  9. владе
  10. orarily

Suppresses

  1. ⏎
  2. ⏎
  3. ⏎⏎
  4. ⏎
  5. ⏎
  6. ⏎⏎
  7. ⏎ ⏎