Neuron Explorer

Qwen3-0.6B

unclear: qualifier words after a noun

A neuron in layer 13 of Qwen3-0.6B. Its strongest snippets are mixed: they peak on different tokens. It is featured, so it has steering outputs.

What this neuron does

This neuron fires on a mix of ordinary words that end or qualify a phrase, such as "out of order", "over-the-counter" and "different people". The pattern is hard to pin down, and steering did not show a clear theme.

Written by Claude from this neuron's activations and steering outputs. Confidence: low. 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

    list the actions from your paragraph. Review the list to make sure the actions are in time order (in the order they happened). Rearrange any actions that are out of order. Make a copy of this Google doc or download a Word template.⏎

  2. #2 peak

    you may be attempting to replicate a study because you do not trust the findings of the original study. A recent example includes the creation of stem cells from ordinary cells, a finding that was so extraodinary that it led several laboratories to quickly try

  3. #3 peak

    of shock management in the trauma patient and the treatment of excessive anticoagulation related to medical treatment are discussed elsewhere. (See "Correcting excess anticoagulation after warfarin" and "Initial evaluation and management of shock in adult

  4. #4 peak

    advice of a medical professional. Consult your medical professional for guidance before changing or undertaking any diet, exercise or health program, or before using any prescription or over-the-counter drug. Neither the author nor Walmart.com

  5. #5 peak

    and identify differences. Exposure to new environments inspires people to look for things that are familiar. Seeking similarities helps people make more connections with others to better understand and accept different people. Traveling kids will see the things<|endoftext|>Report on Macdonald (

  6. #6 peak

    aging.<|endoftext|>After all, Germany has a historic Guilt, Angela Merkel tells a concerned citizen who asks if Merkel does not have a duty to protect Germans from Muslim culture and terrorism! See video further down.⏎Angela Merkel defends German Suicide:

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: the idea that a square has a lot of information about a lot of things about a square... if a square has a lot about a lot about a square... if a lot about a square... if a lot about a square... if a

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

    to lead an active, healthy lifestyle. While seniors may need to eat less than younger individuals, losing your appetite altogether could be a sign of a more serious medical condition.⏎What Causes Loss of Appetite in Seniors?⏎There are a variety of

  2. #2 peak

    depression and psychosis found in the illnesses.⏎Dr. Wilson notes, "I do see a difference in blood sugar levels and depression, but I have not seen a case that controlling

  3. #3 peak

    the body⏎- Specific details about the risks associated with marijuana use, particularly use by youth⏎- The impacts marijuana use has on the brain⏎- Which medical conditions medical marijuana treats⏎- The impact smoking or vaping has on the lungs and other

  4. #4 peak

    but they have also been affected by pollution and river-engineering works.⏎The freshwater pearl mussel was given full legal protection in 1998 but unfortunately illegal activity still continues. Every year we still come across significant ‘kills’ where piles

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. L12:1371 suppresses
  2. L11:327 excites
  3. L12:294 suppresses
  4. L12:529 excites
  5. L12:841 suppresses
  6. L12:395 excites
  7. L11:1864 suppresses
  8. L11:696 suppresses
  9. L4:1140 suppresses
  10. L12:904 suppresses
  11. L0:2770 weakly excites
  12. L12:2201 weakly excites
  13. L12:1991 weakly suppresses
  14. L12:2295 weakly suppresses
  15. L12:428 weakly suppresses
  16. L12:550 weakly suppresses
  17. L8:2885 weakly excites
  18. L2:1895 weakly excites
  19. L12:1071 weakly excites
  20. L11:146 weakly suppresses
  21. L11:1868 weakly excites
  22. L12:358 weakly excites
  23. L12:553 weakly suppresses
  24. L7:231 weakly excites

Feeds later layers

  1. L16:1587 excites
  2. L15:2966 suppresses
  3. L14:2245 weakly suppresses
  4. L17:780 weakly excites
  5. L14:202 weakly excites
  6. L14:2390 weakly excites
  7. L14:899 weakly excites
  8. L15:2662 weakly suppresses
  9. L15:266 weakly excites
  10. L19:1674 weakly suppresses
  11. L14:1900 weakly excites
  12. L18:3033 weakly suppresses
  13. L14:536 weakly suppresses
  14. L17:1293 weakly suppresses
  15. L15:1629 weakly excites
  16. L16:2642 weakly suppresses
  17. L16:1828 weakly excites
  18. L18:1535 weakly suppresses
  19. L20:2264 weakly excites
  20. L15:2223 weakly excites
  21. L14:1801 weakly excites
  22. L14:2041 weakly suppresses
  23. L22:1779 weakly suppresses
  24. L14:1877 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 -2.95 and -2.792 values between -2.46 and -2.291 values between -2.29 and -2.132 values between -2.13 and -1.972 values between -1.97 and -1.802 values between -1.80 and -1.648 values between -1.64 and -1.4711 values between -1.47 and -1.3110 values between -1.31 and -1.1427 values between -1.14 and -0.9836 values between -0.98 and -0.8277 values between -0.82 and -0.65101 values between -0.65 and -0.49180 values between -0.49 and -0.32367 values between -0.32 and -0.16932 values between -0.16 and 0.01858 values between 0.01 and 0.17466 values between 0.17 and 0.34275 values between 0.34 and 0.50173 values between 0.50 and 0.67121 values between 0.67 and 0.8372 values between 0.83 and 0.9962 values between 0.99 and 1.1639 values between 1.16 and 1.3220 values between 1.32 and 1.4921 values between 1.49 and 1.6510 values between 1.65 and 1.8214 values between 1.82 and 1.982 values between 1.98 and 2.147 values between 2.14 and 2.314 values between 2.31 and 2.472 values between 2.47 and 2.641 values between 2.64 and 2.801 values between 2.80 and 2.971 values between 3.46 and 3.62offnegativepositiveoften

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. already
  2. 掖
  3. oken
  4. ominated
  5. already
  6. elle
  7. ске
  8. essen
  9. руб
  10. ibe

Suppresses

  1. orial
  2. quests
  3. peria
  4. strar
  5. `(
  6. ียน
  7. mdir
  8. asons
  9. erals
  10. dress