Neuron Explorer

Qwen3-0.6B

period then line break at end of line

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 weakly on a full stop that is immediately followed by a line break, which usually happens at the end of a sentence or list item in web text. It is a weak signal compared with the strongest neurons. Steering did not produce a clear theme.

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

    80 p. :|⏎|Number of Pages||99|⏎|2||WCC exchange -- 2|⏎nodata File Size: 5MB.⏎Yet for many in the Roman Catholic church in Latin America, not to mention in

  2. #2 peak

    Arabic. Number of native speakers: 221 million. …⏎What is the best app to learn Hebrew?⏎Online Hebrew courses⏎- Learn Hebrew Easily.⏎- Mondly Languages.⏎- i<|endoftext|>As the citizens of the United States follow

  3. #3 peak

    AND BENEFICIATION PRACTICES . In most other cases, the ore is crushed and/or ground, prior to leaching or flotation. Each of the.⏎10 OPERATIONS: ORE PROCESSING - Midas Gold⏎Ore

  4. #4 peak

    decreased by nearly two-thirds.⏎“Tsunami Early Warning System Moves Ahead at UN Workshop,” UN News Service, 14 June 2006.⏎Link: http://www.un.org/apps/news/story.asp?NewsID=1

  5. #5 peak

    peak of our region’s fire emergency, somebody rang our Rural Fire Service (RFS) headquarters and complained: ‘Why haven’t you put the fires out?’.⏎There are parallels between the current pandemic, the recent fire emergency and ‘wildfires

  6. #6 peak

    the Murray B. Sachs Professor in the Department of Biomedical Engineering at The Johns Hopkins University Schools of Engineering and Medicine and the project’s principal investigator… Continue reading.⏎Natalia Trayanova, a professor

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: Is it possible for thesitiess tos

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

    its vestiges.⏎Long before the liberal revolution of Europe dealt the most effective political blows against feudal power in the 17th⏎and 18th century, a cultural revolution took shape in the Renaissance which

  2. #2 peak

    air in buildings. Its use involves the construction of vertical ventilation channels, which are designed to carry away warm air from the upper layers of the interior to the outside⏎Its operation is based on the pressure difference generated in the ventilation duct under appropriate atmospheric

  3. #3 peak

    and rocket projectiles respectively.⏎Mahogany Wood. Wingspan 19 inches, Length 19 inches.<|endoftext|>A random variable X has the cumulative distribution function⏎Calculate the variance of X.⏎The answer specifies that the density function is⏎I

  4. #4 peak

    It was named after George C. Marshall, who was the US Secretary of State. Thousands of millions of dollars were provided for rebuilding cities, roads, industries etc⏎Definition from the Longman Dictionary of Contemporary English Advanced Learner's Dictionary.⏎Dictionary

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. L17:829 suppresses
  2. L10:84 suppresses
  3. L8:32 excites
  4. L18:429 suppresses
  5. L19:187 suppresses
  6. L16:359 suppresses
  7. L13:10 excites
  8. L10:33 suppresses
  9. L15:123 suppresses
  10. L21:173 excites
  11. L1:1 excites
  12. L16:170 suppresses
  13. L18:5 suppresses
  14. L6:67 excites
  15. L9:25 excites
  16. L18:227 suppresses
  17. L21:130 excites
  18. L14:56 suppresses document ends and list line breaks
  19. L5:331 suppresses
  20. L18:48 weakly suppresses
  21. L18:124 weakly suppresses
  22. L6:4 weakly suppresses
  23. L9:7 weakly excites
  24. L14:57 weakly suppresses

Feeds later layers

  1. L25:577 strongly excites
  2. L23:393 suppresses
  3. L27:2408 suppresses
  4. L24:1512 excites
  5. L26:1008 suppresses
  6. L25:535 excites commas inside number lists
  7. L24:318 suppresses
  8. L24:2198 excites closing parentheses in lists
  9. L24:2281 weakly excites
  10. L26:362 weakly excites
  11. L27:740 weakly excites
  12. L26:505 weakly excites unclear: opening quote marks after a line break
  13. L26:402 weakly suppresses
  14. L25:396 weakly suppresses
  15. L23:1613 weakly suppresses
  16. L23:585 weakly suppresses
  17. L23:703 weakly suppresses
  18. L25:2123 weakly excites
  19. L27:2630 weakly excites
  20. L27:1007 weakly excites
  21. L27:2268 weakly excites
  22. L23:806 weakly suppresses
  23. L26:47 weakly suppresses
  24. L26:2753 weakly 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 -29.13 and -28.012 values between -26.89 and -25.771 values between -25.77 and -24.651 values between -22.41 and -21.292 values between -21.29 and -20.171 values between -20.17 and -19.052 values between -19.05 and -17.933 values between -17.93 and -16.814 values between -16.81 and -15.693 values between -14.57 and -13.451 values between -12.33 and -11.214 values between -10.09 and -8.976 values between -8.97 and -7.851 values between -7.85 and -6.734 values between -6.73 and -5.619 values between -5.61 and -4.4919 values between -4.49 and -3.3755 values between -3.37 and -2.25183 values between -2.25 and -1.131380 values between -1.13 and -0.011998 values between -0.01 and 1.11140 values between 1.11 and 2.2334 values between 2.23 and 3.3512 values between 3.35 and 4.475 values between 4.47 and 5.595 values between 5.59 and 6.713 values between 6.71 and 7.831 values between 7.83 and 8.957 values between 8.95 and 10.077 values between 10.07 and 11.195 values between 11.19 and 12.317 values between 12.31 and 13.431 values between 13.43 and 14.551 values between 14.55 and 15.67offnegativepositiveoften

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. elda
  2. TING
  3. living
  4. ��
  5. ught
  6. living
  7. 脔
  8. world
  9. BED
  10. quito

Suppresses

  1. s
  2. sar
  3. sil
  4. ات
  5. 员
  6. sie
  7. enticate
  8. acters
  9. sand
  10. sse