Neuron Explorer

Qwen3-0.6B

mid-sentence line breaks

A neuron in layer 21 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 line-break characters that appear in the middle of a sentence, where the text was wrapped onto a new line. Turning it on made the model's text drift into unrelated topics, so the steering does not clearly show the pattern.

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

    else, themselves included.⏎Census-takers designated them Russian-born, and they were, literally.⏎Geographically they were as well, since they originated from at⏎least six distinct nationalities within the vast Russian Empire.⏎In every other respect,

  2. #2 peak

    CD.⏎High quality scanned images of the whole of the original book. This CD has been bookmarked for easy⏎navigation, and pages can be searched, browsed, enlarged and printed out if

  3. #3 peak

    uffed at the knees, a style that⏎involves folding the garment in half and tucking it under the waist. This style⏎of draping the⏎dhoti is known as the half-fold. The dhoti can be

  4. #4 peak

    âyâna) is most commonly known in the Chinese Buddhist literature by the name of Açvaghosha. But according⏎to his Life he was also called Kung-tê-jih (i.e

  5. #5 peak

    from organic matter as opposed to fossil fuels.⏎“As a renewable source of natural gas, the city’s sewage will offer a steady and sustainable supply of clean fuel well⏎into the future,” said Michael Jordan, Environmental Services’ director. “The filling

  6. #6 peak

    teaser just like this one...⏎Or, just get a random brain teaser⏎If you become a registered user you can vote on this brain teaser, keep track of⏎which ones you have seen, and even make your own.⏎Back to Top<|endoftext|>

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: I will see, as I will see, the shape of the shape. To see the shape of the shape. That will mean, I think, that the shape is the meaning, to be the shape, to be the meaning. that

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

    How do land use practices and management efforts interact with natural disturbances and climatological cycles to affect estuarine water quality and suitability for ecologically (and economically) important species?<|endoftext|>Not Just for Kids⏎The

  2. #2 peak

    to increased pressure inside the skull, including:⏎Brain herniation can occur:⏎Patients with a brain herniation have:⏎A brain and nervous system (neurological) exam shows changes in alertness (consciousness). Depending on the severity of the

  3. #3 peak

    ology Course Description⏎- 2 Case Studies⏎- 3 Outbreak Investigations⏎- 4 Clinical Trials⏎- 5 Cohort (Longitudinal) Studies⏎- 6 Case-Control Studies⏎- 7 Mesures of Mor

  4. #4 peak

    parallel to the sun (for example, sunrise striking a horizontal rooftop), the angle of incidence is 90°.⏎Array — see photovoltaic (PV) array.⏎Battery Cell — The simplest operating unit in a storage battery. It consists

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. L20:564 strongly suppresses
  2. L20:1147 strongly suppresses comma after a sentence subject
  3. L20:281 strongly excites
  4. L16:2076 strongly suppresses
  5. L20:1447 strongly suppresses
  6. L16:1323 suppresses
  7. L16:1258 suppresses
  8. L17:310 suppresses
  9. L16:2315 excites
  10. L16:2983 excites
  11. L18:933 suppresses
  12. L20:1595 suppresses
  13. L20:1781 suppresses
  14. L20:526 suppresses
  15. L20:90 excites
  16. L9:1515 excites line break splitting a phrase
  17. L17:1192 suppresses
  18. L16:1003 excites
  19. L16:1608 suppresses
  20. L18:237 suppresses
  21. L15:2449 excites line breaks mid-sentence
  22. L17:359 suppresses
  23. L18:879 excites
  24. L16:1246 suppresses

Feeds later layers

  1. L27:3020 weakly excites
  2. L26:2128 weakly excites decimal points inside numbers
  3. L26:2731 weakly excites
  4. L27:2135 weakly excites
  5. L26:747 weakly excites
  6. L23:1624 weakly suppresses
  7. L25:1279 weakly excites
  8. L26:148 weakly excites
  9. L24:662 weakly suppresses
  10. L25:387 weakly excites
  11. L27:362 weakly excites
  12. L27:2258 weakly excites
  13. L22:2089 weakly suppresses
  14. L23:1185 weakly excites
  15. L27:1354 weakly excites
  16. L25:645 weakly excites
  17. L23:334 weakly excites
  18. L27:369 weakly excites
  19. L25:969 weakly excites
  20. L25:1736 weakly excites
  21. L23:87 weakly excites
  22. L26:2634 weakly excites
  23. L25:808 weakly excites
  24. L25:2869 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 -13.52 and -12.862 values between -10.21 and -9.551 values between -8.88 and -8.221 values between -8.22 and -7.564 values between -7.56 and -6.901 values between -6.90 and -6.236 values between -6.23 and -5.5713 values between -5.57 and -4.918 values between -4.91 and -4.2411 values between -4.24 and -3.5830 values between -3.58 and -2.9245 values between -2.92 and -2.2593 values between -2.25 and -1.59220 values between -1.59 and -0.93614 values between -0.93 and -0.272334 values between -0.27 and 0.40335 values between 0.40 and 1.0693 values between 1.06 and 1.7246 values between 1.72 and 2.3918 values between 2.39 and 3.056 values between 3.05 and 3.716 values between 3.71 and 4.386 values between 4.38 and 5.045 values between 5.04 and 5.701 values between 5.70 and 6.364 values between 6.36 and 7.031 values between 7.03 and 7.691 values between 8.35 and 9.021 values between 10.34 and 11.001 values between 12.33 and 12.99offnegativepositiveoften

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. rish
  2. же
  3. справ
  4. łu
  5. ện
  6. ὕ
  7. ích
  8. 下来
  9. �
  10. edio

Suppresses

  1. /etc
  2. AFC
  3. alike
  4. Traditional
  5. ロン
  6. imulation
  7. conversion
  8. 틱
  9. Alicia
  10. common