Neuron Explorer

Qwen3-0.6B

the number '100' in numerals

A neuron in layer 2 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 "0" in the number 100, such as "100 years" or "10,000". Turning it up did not produce a consistent change in what the model wrote.

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

    the biggest collection of Educational Titanic resources and should be the starting point for any teacher and student. Their education section is very comprehensive.⏎National Geographic Titanic 100: A very current and fresh collection of Titanic content with a large amount of input

  2. #2 peak

    CCD reports.⏎Yet another factor is the ISO rating of your CCD. At ISO 400 you'll notice more hot pixels than at ISO 100, simply because the signal is amplified.⏎A hot pixel can range from bright white

  3. #3 peak

    backs, stock prices. One bizarre rumor about the AIDS virus was that it was developed in a Western laboratory and that the World Health Organization inoculated 100,000

  4. #4 peak

    along a 100-kilometer long north-south linear vent system over the past 5 million

  5. #5 peak

    function was checked again during four assessments over 11 years using the Modified Mini-Mental State examination (3MS), which is graded on a 100 point scale. A DASH diet adherence score was created based on consumption levels of

  6. #6 peak

    AbstractThe tobacco epidemic is one of the major global public health problems today that is being facilitated through a variety of complex factors. The tobacco epidemic killed 100 million people worldwide in the 20th century and it

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 a lot of power, and I have a lot of power." So, I have a lot of power. What is the result of the first experiment, and what is the second thought? So, both sides of the first

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

    such problems could prevent the potential development and perpetuation of psychological disorders among these individuals that may affect their efforts to address their obesity,” study researcher Marzieh Hosseini Araghi, BSc, MPH, of the University of Birmingham,

  2. #2 peak

    instance, "hok" warns of serious aerial threats -- usually eagles -- whereas "hok-

  3. #3 peak

    of 15,795 razorback sucker and 11,696 bonytail were released with 134.2 kilohertz (kHz)

  4. #4 peak

    ives, in areas of skin away from the bite.⏎• Hoarseness or shortness of breath.⏎Anaphylaxis requires emergency medical attention. People

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. L0:697 excites the last zero in round numbers
  2. L0:920 weakly excites digit '9' inside years and numbers
  3. L1:288 weakly excites
  4. L1:2330 weakly excites
  5. L1:178 weakly suppresses
  6. L1:2310 weakly excites
  7. L0:1741 weakly suppresses
  8. L1:1900 weakly excites
  9. L1:544 weakly excites
  10. L0:1541 weakly excites
  11. L0:369 weakly excites
  12. L1:565 weakly excites
  13. L0:529 weakly suppresses
  14. L0:1121 weakly excites
  15. L1:3022 weakly suppresses
  16. L1:293 weakly suppresses
  17. L1:2772 weakly excites
  18. L1:2884 weakly excites
  19. L1:2414 weakly excites
  20. L0:85 weakly suppresses
  21. L1:50 weakly excites
  22. L1:2385 weakly excites
  23. L0:2594 weakly excites
  24. L0:2108 weakly suppresses

Feeds later layers

  1. L23:1099 suppresses
  2. L20:117 weakly excites
  3. L8:2995 weakly suppresses
  4. L23:869 weakly excites
  5. L8:3067 weakly suppresses
  6. L7:2031 weakly suppresses
  7. L7:2098 weakly excites
  8. L7:2541 weakly suppresses
  9. L22:1509 weakly excites
  10. L18:2231 weakly suppresses
  11. L4:505 weakly suppresses
  12. L20:145 weakly excites
  13. L5:73 weakly suppresses
  14. L10:424 weakly suppresses
  15. L17:410 weakly excites
  16. L19:1851 weakly excites
  17. L10:813 weakly excites
  18. L26:859 weakly excites
  19. L9:307 weakly suppresses
  20. L23:2138 weakly suppresses
  21. L25:1310 weakly excites
  22. L6:715 weakly excites
  23. L14:2954 weakly excites
  24. L12:1909 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.

2 values between -0.99 and -0.923 values between -0.64 and -0.573 values between -0.57 and -0.494 values between -0.49 and -0.428 values between -0.42 and -0.3511 values between -0.35 and -0.2838 values between -0.28 and -0.21165 values between -0.21 and -0.14417 values between -0.14 and -0.071185 values between -0.07 and 0.001247 values between 0.00 and 0.08531 values between 0.08 and 0.15191 values between 0.15 and 0.2247 values between 0.22 and 0.2918 values between 0.29 and 0.369 values between 0.36 and 0.436 values between 0.43 and 0.505 values between 0.50 and 0.574 values between 0.57 and 0.653 values between 0.65 and 0.722 values between 0.72 and 0.792 values between 0.79 and 0.862 values between 1.00 and 1.071 values between 1.07 and 1.141 values between 1.14 and 1.221 values between 1.71 and 1.792 values between 1.79 and 1.86offnegativepositiveoften

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. izzie
  2. 让用户
  3. śni
  4. <tag
  5. orama
  6. .Companion
  7. ousse
  8. urance
  9. ushi
  10. [tag

Suppresses

  1. �
  2. press
  3. -offs
  4. reserv
  5. pair
  6. ACT
  7. significant
  8. �
  9. relocation
  10. olumes