Neuron Explorer

Qwen3-0.6B

birth dates, the day before the year

A neuron in layer 10 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 spot right before a birth year, just after a day and month like 'born December 7,'. It is one of the clearest patterns here. Turning it up made the writing drift into unrelated topics.

Written by Claude from this neuron's activations and steering outputs. Confidence: high. Think it is wrong? Flag it (opens an email to the maker with the neuron already filled in).

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

    full Titus Vespasianus Augustus, original name Titus Flavius Vespasianus, (born Dec. 30, 39 ce—died Sept.

  2. #2 peak

    , who helped to promote and publish his work.⏎What makes Robert Frost unique?⏎Robert Frost, in full Robert Lee Frost, (born March 26, 1874, San Francisco, California, U.S.—died January

  3. #3 peak

    during the final manned flight of the Skylab space station in 1973 and 1974.⏎He was born January 23, 1930, in Okemah, Oklahoma. He received a Bachelor

  4. #4 peak

    Ticknor⏎Curtis, George Ticknor, American writer on legal topics, was born at Watertown, Mass., Nov. 28, 1812; and died at New York,

  5. #5 peak

    |⏎|Born||December 7, 1928⏎|Era||20th / 21

  6. #6 peak

    was born in Montpelier, Vermont, on December 26, 1837. A hero of the Civil War and Spanish-American War,

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. A company has a report that is not in the system. 2. A company has a report that is not in the system. 2. A company has a report that is not in the system. 3.

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

    . AROMATIC COMPOUNDS ARE THOSE ORGANIC COMPOUNDS WHICH HAVE BENZENE RING (CYCLOHEXA-1,3,5-TRIENE). AROMATIC COMPOUNDS

  2. #2 peak

    and one family name, but can have as many additional given names and middle names as they like. A middle name is usually a recognised surname and not a patronymic.⏎What are the () called?⏎() – This are simply termed as brackets

  3. #3 peak

    : C F O S U.⏎No direct anagrams for focus found in this word list.⏎List shorter words within focus, sorted by length⏎All words formed from focus by changing one letter⏎Browse words starting with focus by next letter<|endoftext|>I

  4. #4 peak

    led the way again in Swiss Ball manufacture by eliminating all phthalates from its products.” says Bradley Wilson, Managing Director of AOK Health Pty Ltd.⏎Phthalates are a group of chemicals used to make plastics more flexible and harder to break

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. L6:1894 excites
  2. L9:3038 excites
  3. L6:214 suppresses
  4. L8:590 excites
  5. L8:1945 suppresses
  6. L9:345 excites
  7. L9:2891 suppresses
  8. L4:2290 suppresses
  9. L8:523 suppresses
  10. L6:2498 suppresses
  11. L6:2198 excites
  12. L9:652 suppresses
  13. L8:2252 excites
  14. L6:1047 excites
  15. L6:1783 excites
  16. L7:577 suppresses
  17. L9:2498 suppresses
  18. L8:2919 excites
  19. L7:1121 excites
  20. L9:271 excites
  21. L9:2676 suppresses
  22. L8:2121 excites
  23. L9:1725 suppresses
  24. L7:2340 weakly suppresses

Feeds later layers

  1. L16:2377 excites glosses, translations and unit conversions
  2. L17:958 excites
  3. L12:597 excites
  4. L20:2867 excites
  5. L11:980 weakly excites
  6. L21:2797 weakly excites
  7. L14:2686 weakly excites
  8. L20:2591 weakly excites
  9. L11:1741 weakly excites
  10. L11:1703 weakly excites
  11. L12:1279 weakly excites
  12. L21:2125 weakly suppresses
  13. L11:847 weakly excites
  14. L19:2661 weakly excites
  15. L20:1256 weakly excites
  16. L14:2058 weakly suppresses
  17. L11:2670 weakly excites
  18. L11:2754 weakly excites
  19. L19:1876 weakly excites
  20. L20:1802 weakly excites
  21. L19:1795 weakly excites
  22. L21:2875 weakly excites
  23. L11:1844 weakly excites
  24. L25:1648 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 -3.15 and -2.792 values between -1.70 and -1.348 values between -1.34 and -0.9848 values between -0.98 and -0.62771 values between -0.62 and -0.262743 values between -0.26 and 0.10223 values between 0.10 and 0.4644 values between 0.46 and 0.8217 values between 0.82 and 1.1814 values between 1.18 and 1.5411 values between 1.54 and 1.908 values between 1.90 and 2.264 values between 2.26 and 2.632 values between 2.63 and 2.991 values between 2.99 and 3.353 values between 3.35 and 3.711 values between 3.71 and 4.071 values between 4.07 and 4.431 values between 4.79 and 5.151 values between 5.15 and 5.511 values between 5.87 and 6.231 values between 7.31 and 7.671 values between 10.20 and 10.561 values between 10.92 and 11.28offnegativepositiveoften

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. ños
  2. ancestry
  3. 世家
  4. 姓
  5. thổ
  6. ño
  7. 遗传
  8. 天赋
  9. 天生
  10. born

Suppresses

  1. urg
  2. $action
  3. ]*
  4. UFC
  5. whel
  6. Massage
  7. eu
  8. action
  9. cn
  10. 愦