Neuron Explorer

Qwen3-0.6B

capital W starting a word

A neuron in layer 27 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

Fires on a capital W that starts a word or abbreviation, like Worms, Waste or WBMP. Turning it up pushed the model to repeat the word 'laws', a loose echo of the law-related words this neuron tends to push toward.

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

    also full of nutrients. Some of these nutrients are very valuable additions for your garden soil or compost pile. This is important when it comes to worm castings.⏎Worms do not eat dog waste directly. They eat the bacteria that break down

  2. #2 peak

    title, visit your library in the app or on the desktop website.⏎Chronicles of the imagination Geographical⏎- Walta Evans "School Librarian"<|endoftext|>Wing Tsun or Wing Chun?⏎or Wing Chung or Wing Tjun or V

  3. #3 peak

    course can also be counted as a 1/4 credit in American history.⏎Read with the Best (RWTB) coordinates with Write with the Best (WWTB) (by the same author, Jill Dixon). Volume 1 is

  4. #4 peak

    al., 2011). Paper wasps are likely to influence many other species in terrestrial ecosystems because they are voracious predators of invertebrates (Ward & Ramón-Laca, 2014). Four species of

  5. #5 peak

    A comprehensive approach to improving the lives of waste pickers should assist them and their families to achieve their potential and bring a positive socio-economic change in their lives.⏎Waste pickers

  6. #6 peak

    of Missions for the Pacific, located in San Francisco (WBMP), and the Woman's Board of Missions for the Pacific Islands, located in Hawaii (WBMPI). The three mainland boards cooperated with the maleled American Board of

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: whether le le le le le le le le le le le le le le le le le le le le le le le le le le le le le le le le le le le le le le le le le le le le le le le

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

    communities.⏎- Urban Forests: Urban Releaf and Friends of the Urban Forest are two grassroots organizations that beautify our neighborhoods by planting trees.⏎- Aquatic/Wetland Habitats: An inspiring example of inter-jurisdictional collaboration between

  2. #2 peak

    Historical Society, Scarborough Middle School, and Scarborough Public Library. Exhibits include the marsh, transportation and roads, shipyards and shipwrecks, clamming and lobstering, famous residents, and education.<|endoftext|>

  3. #3 peak

    , Captain Laudonniere states: “Having reconnoitered the river, I landed to talk […]⏎AN ARTIST WHO ACCOMPANIED THE FRENCH EXPEDITION TO FLORIDA UNDER LAUDONNIERE,

  4. #4 peak

    (S.W.), the New England Confederation and the Fur Trade Dutch/French (N.Y. &N.F.) is the third supporting idea. *

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. L22:2646 excites
  2. L23:2441 excites
  3. L23:1906 suppresses
  4. L23:2073 excites
  5. L20:417 excites
  6. L22:369 excites
  7. L21:463 suppresses
  8. L18:533 suppresses
  9. L22:2236 excites
  10. L18:437 weakly excites
  11. L26:1463 weakly excites
  12. L16:74 weakly excites
  13. L21:2605 weakly excites
  14. L10:941 weakly suppresses
  15. L8:1982 weakly excites
  16. L2:1503 weakly excites
  17. L19:2719 weakly excites
  18. L8:14 weakly excites
  19. L26:1564 weakly excites
  20. L26:2456 weakly excites
  21. L16:71 weakly excites
  22. L20:1877 weakly excites
  23. L20:336 weakly excites
  24. L9:1588 weakly excites

Feeds later layers

None: this is the last layer.

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 -32.56 and -31.331 values between -26.38 and -25.141 values between -25.14 and -23.901 values between -22.66 and -21.431 values between -18.95 and -17.711 values between -15.24 and -14.001 values between -12.76 and -11.531 values between -11.53 and -10.292 values between -10.29 and -9.051 values between -9.05 and -7.812 values between -7.81 and -6.582 values between -6.58 and -5.341 values between -5.34 and -4.104 values between -4.10 and -2.8613 values between -2.86 and -1.63177 values between -1.63 and -0.393631 values between -0.39 and 0.8558 values between 0.85 and 2.094 values between 2.09 and 3.332 values between 3.33 and 4.561 values between 4.56 and 5.801 values between 13.23 and 14.461 values between 15.70 and 16.94offnegativepositiveoften

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. laws
  2. law
  3. laws
  4. Laws
  5. legislation
  6. Law
  7. -law
  8. 定律
  9. 规律
  10. law

Suppresses

  1. inal
  2. inals
  3. edBy
  4. INAL
  5. istics
  6. acles
  7. ellaneous
  8. setDescription
  9. ativo
  10. rapper