Neuron Explorer

Qwen3-0.6B

'by' giving the amount of a difference

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 lights up on the word "by" when it introduces how much something changed or differed, as in "predates the Altair by almost two years" or "shifted in the same direction by the same amount." When turned up, the model's writing tended toward numbers and amounts, which fits this pattern.

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

    well as at the fueling station, which also is being built and operated in partnership with NW Natural.⏎Trucks running on natural gas cut smog-producing pollutants by up

  2. #2 peak

    wrecks. At nearly 3,000 years old, this Bronze Age find is particularly significant, not just because it predates the birth of Christ by a millennium, but because in addition to the weapons and the precious gold jewelry,

  3. #3 peak

    recording of a human voice.⏎The recording played Thursday predates Thomas Edison's invention of the phonograph (previously thought to have recorded the first sound) by 17 years. It captured about 10 seconds of the French folks

  4. #4 peak

    eter. If the voltmeter is not correctly zeroed before use, then all values measured by the voltmeter will be biased, i.e., offset by some constant amount or factor. However, even if the utmost care is taken in

  5. #5 peak

    employed the Intel 4004 CPU, were the first American microcomputers ever to be sold and pre-date the Altair 8800 by almost two years.⏎The Intellec Microcomputer Development Systems (MDS)

  6. #6 peak

    characters entered by the user, but in reverse order. The second string contains the 5 characters entered by the user, but whose ASCII code has been decremented 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: the question that I think I can be a person who is as strong as the strongest person who is the strongest person who is... the strongest one who... the strongest one who... the strongest one... the strongest. ... the strongest... 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

    patient.⏎Medical-Surgical Nursing at a Glance is ideal for nursing students looking for a clear and visual summary of care of the adult patient.⏎Results 1-5 of 41⏎The nurse practitioner is skilled at

  2. #2 peak

    (6-8) Show all Grade Level⏎Resource Type: Activities⏎Results 1 - 7 of 7 matches⏎From Grid to Home⏎Lisa Gardiner, National

  3. #3 peak

    Necessary intensities are in the range between 1010- 1011W/cm2.1, 2 at such intensities, gas molecules

  4. #4 peak

    V | W | X | Y | Z | 0-9⏎21 results for Schools: |⏎See previous results⏎See results 11 - 15⏎View all

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. L12:2545 suppresses
  2. L6:756 suppresses
  3. L10:868 suppresses
  4. L10:1214 suppresses
  5. L13:884 excites
  6. L5:685 suppresses
  7. L13:1308 excites
  8. L6:1452 weakly suppresses
  9. L3:2093 weakly excites
  10. L7:1088 weakly suppresses
  11. L12:1777 weakly suppresses
  12. L13:513 weakly excites
  13. L13:2770 weakly suppresses
  14. L12:2801 weakly excites percentages and time spans in statistics
  15. L0:2932 weakly suppresses
  16. L9:1108 weakly suppresses
  17. L11:2493 weakly suppresses
  18. L13:2915 weakly suppresses
  19. L9:2078 weakly suppresses
  20. L2:1045 weakly excites
  21. L4:1672 weakly suppresses
  22. L12:740 weakly excites
  23. L11:95 weakly excites
  24. L8:2164 weakly excites

Feeds later layers

  1. L18:3014 strongly excites
  2. L16:3022 excites
  3. L17:2721 excites
  4. L16:2032 excites
  5. L15:2091 excites
  6. L15:3028 excites
  7. L24:722 excites 'about' or 'approximately' before a number
  8. L24:587 excites
  9. L15:1909 excites blank slot before a number
  10. L19:1419 excites
  11. L17:222 excites
  12. L17:1951 excites
  13. L21:175 excites
  14. L17:1570 excites
  15. L17:1804 excites
  16. L17:1504 excites
  17. L18:291 suppresses
  18. L17:2952 excites
  19. L17:2208 excites
  20. L17:1512 excites
  21. L16:1775 excites
  22. L18:2976 excites
  23. L19:1458 excites
  24. L19:2151 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.

8 values between -0.84 and -0.6648 values between -0.66 and -0.49297 values between -0.49 and -0.311015 values between -0.31 and -0.131664 values between -0.13 and 0.04713 values between 0.04 and 0.22132 values between 0.22 and 0.3919 values between 0.39 and 0.575 values between 0.57 and 0.741 values between 0.74 and 0.921 values between 0.92 and 1.101 values between 1.45 and 1.622 values between 1.62 and 1.801 values between 4.79 and 4.961 values between 6.02 and 6.19offnegativepositiveoften

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. 一刀
  2. 一秒
  3. 几分钟
  4. 一个小时
  5. 多久
  6. 半个
  7. few
  8. 一点点
  9. 一步
  10. -half

Suppresses

  1. sville
  2. urai
  3. jec
  4. juxtap
  5. uß
  6. surrounded
  7. antasy
  8. 周三
  9. ypad
  10. sounded