Neuron Explorer

Qwen3-0.6B

comma in dates and birth lines

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 the comma that separates the day from the year in dates, such as "September 24, 1896" or "March 26, 1874", mostly in biographies and reference entries. Its negative side is very weak. Steering results were scrambled or off topic, so they do not show what it does.

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

    – in contrast to chemical pesticides – have no adverse effects on animals or humans.<|endoftext|>F. Scott Fitzgerald⏎Francis Scott Key Fitzgerald (September 24, 1896 – December 21,1940)

  2. #2 peak

    of the third week?⏎3. Solve for the following.⏎a. In a geometric sequence, and . Find .⏎b. In a geometric sequence, and . Find .⏎In Example 1b and Review Questions 5 and 6, students

  3. #3 peak

    panels can be installed on rooftops or on the ground.⏎- Sunpower solar panels<|endoftext|>By Mortimer J. Adler⏎Aristotle (384 - 322 B.C.) taught common sense to Alexander the nice and,

  4. #4 peak

    Pound, 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

  5. #5 peak

    Russia|⏎|-||Declared1||November 18, 1918|⏎|-||Recognized||January 26, 1921|⏎|-||Soviet occupation||August 5

  6. #6 peak

    non-alignment and of socialism grounded in Islam.⏎Mohammad Hatta was born in Bukittinggi, West Sumatra, Indonesia, on August 12, 1902. Although his father died while he was an infant,

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:ThetransformationalimpactofDigitalSolutionsintotheBusinessModelofTraditionalBusinessModelofDigital-Platform-Applications.在Iamamostofamaintainingthetraditionalbusinessmodelofdigital-priv-priv-priv-priv

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

    and alternating current in the series RLC circuit at an instant are given by⏎v=Vm sin ωt and i = I m sin( ω t + φ)⏎where ϕ is the phase angle between υ and i.

  2. #2 peak

    .<|endoftext|>Start a 10-Day Free Trial to Unlock the Full Review⏎Why Lesson Planet?⏎Find quality lesson planning resources, fast!⏎Share & remix collections to

  3. #3 peak

    the same amount of time running at a pace of 5 miles per hour. A study in the May 1996 issue of the "Journal of the American Medical Association" also concluded that running on a treadmill is superior to using the

  4. #4 peak

    significant.⏎“Even total mortality went in the right direction,” said Gaziano, whose new data comes from the Physicians Health Study II published Wednesday in the Journal of the American

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. L19:963 strongly suppresses
  2. L18:246 strongly suppresses
  3. L17:15 strongly suppresses
  4. L20:59 strongly excites
  5. L19:763 strongly suppresses
  6. L14:140 strongly excites
  7. L19:1133 strongly suppresses last word before end of document
  8. L19:430 strongly suppresses
  9. L19:1174 strongly suppresses
  10. L19:1172 strongly suppresses
  11. L20:433 strongly excites
  12. L19:1005 strongly suppresses
  13. L20:193 strongly suppresses
  14. L14:309 strongly suppresses
  15. L21:164 strongly suppresses
  16. L16:723 strongly excites
  17. L18:2497 strongly suppresses hyphen inside compound words
  18. L6:42 strongly excites
  19. L4:2129 strongly excites
  20. L21:214 strongly excites
  21. L20:489 strongly suppresses
  22. L19:674 strongly suppresses
  23. L19:2177 strongly excites
  24. L19:981 strongly suppresses

Feeds later layers

  1. L27:33 strongly suppresses
  2. L24:39 strongly suppresses
  3. L26:21 strongly excites
  4. L25:109 strongly excites
  5. L25:161 strongly excites
  6. L24:51 strongly suppresses
  7. L25:30 strongly suppresses
  8. L27:263 strongly excites
  9. L27:71 strongly excites
  10. L27:112 strongly suppresses
  11. L23:2741 strongly excites
  12. L23:80 strongly suppresses blank slot before a year or number
  13. L27:97 strongly suppresses
  14. L27:117 strongly suppresses
  15. L27:825 excites
  16. L25:311 excites
  17. L23:60 suppresses
  18. L27:85 suppresses
  19. L27:11 excites unclear: mixed fragments in lists and pronunciations
  20. L27:126 suppresses
  21. L26:189 suppresses
  22. L24:609 excites
  23. L24:33 excites
  24. L25:112 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.

59 values between -3.19 and -2.11217 values between -2.11 and -1.02223 values between -1.02 and 0.06210 values between 0.06 and 1.14238 values between 1.14 and 2.23278 values between 2.23 and 3.31314 values between 3.31 and 4.39367 values between 4.39 and 5.47353 values between 5.47 and 6.56334 values between 6.56 and 7.64290 values between 7.64 and 8.72275 values between 8.72 and 9.81201 values between 9.81 and 10.89176 values between 10.89 and 11.97115 values between 11.97 and 13.0581 values between 13.05 and 14.1442 values between 14.14 and 15.2249 values between 15.22 and 16.3023 values between 16.30 and 17.3915 values between 17.39 and 18.479 values between 18.47 and 19.5513 values between 19.55 and 20.639 values between 20.63 and 21.725 values between 21.72 and 22.805 values between 22.80 and 23.882 values between 23.88 and 24.971 values between 24.97 and 26.051 values between 28.21 and 29.301 values between 31.46 and 32.551 values between 33.63 and 34.711 values between 39.04 and 40.13offnegativepositiveoften

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. ,
  3. ),
  4. ,
  5. ).
  6. )
  7. 。
  8. .
  9. )(
  10. ):⏎

Suppresses

  1. 的作用
  2. 的工作
  3. *self
  4. 的对象
  5. sar
  6. 'S
  7. 的颜色
  8. Cx
  9. "She
  10. 幸运