Neuron Explorer

Qwen3-0.6B

opening bracket before an acronym

A neuron in layer 26 of Qwen3-0.6B. Its strongest snippets are fairly consistent: several of them peak on the same token. It is featured, so it has steering outputs.

What this neuron does

This neuron fires on the opening bracket that comes right after a full name, just before its short form, such as the bracket before "ACA" or "AAP". It goes quiet, and sometimes slightly negative, on opening quote marks. Turning it down made the text fill with blank placeholders, and turning it up gave ordinary rambling, so steering shows no 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

    US citizenship were also three times less likely to have private or public health insurance, resulting in minimal to no access to comprehensive reproductive health care. The Affordable Care Act (ACA) preserved a 1996 restriction requiring non-citizens to

  2. #2 peak

    choice of format and writing style.⏎English 102. It is assumed that you already have basic knowledge and skill in critical thinking, library research, MLA (Modern Language Association) documentation, American English grammar, style, usage, and persuasive

  3. #3 peak

    published books and computational analyses of social media data - to disentangle the dynamic relationship between Japanese public opinion and Japan–South Korea relations.⏎

  4. #4 peak

    in the individual market. These plans can be purchased directly from insurance providers or through state-based exchanges, such as the Health Insurance Marketplace established under the Affordable Care Act (ACA).⏎c. Government Programs: The US government provides medical insurance through programs like

  5. #5 peak

    shift in patient care. Advances in endoscopic techniques and in imaging help achieve these goals. Among endoscopy techniques are:⏎- Endobronchial ultrasound (EBUS)⏎- Electromagnetic navigation (EMN) with volumetric high

  6. #6 peak

    32K code size limit; you can break up your code into a main portion (of at most 32K), and a number of libraries (each of at most 32K).⏎GLib (pronounced ``gee

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: __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __

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

    sharp phrases, interesting metaphors, rhyming and code-switching. Calò words and phrases, such as “la neta,” (the truth) “trucha,” (vigilant)

  2. #2 peak

    lion had an important role in Roman mythology, since it was associated with the god Hercules. Notoriously, lions were integral to the form of capital punishment known as ‘damnatio ad bestias’, whereby condemned criminals were pitted against

  3. #3 peak

    Denudate [dih-NOO-deyt ](verb): to make bare; strip; denude.⏎This fun word comes from the latin word “denudatus” which means “naked” or “ bare.”⏎With the

  4. #4 peak

    auxilliary forces rushed to the rescue of the farmer, o'erwhelming the⏎enemy and with one fell swoop, bringing on them consternation and⏎ruin dire. It appears that the foe, or their progenitors,⏎had

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. L25:239 strongly excites
  2. L20:563 excites
  3. L19:509 excites
  4. L22:65 suppresses
  5. L21:321 excites
  6. L21:29 excites
  7. L19:124 suppresses
  8. L24:77 excites opening parenthesis after a name
  9. L23:80 excites blank slot before a year or number
  10. L19:1207 suppresses
  11. L20:1585 excites
  12. L25:222 suppresses
  13. L19:1174 excites
  14. L18:76 suppresses blank spot before a number
  15. L23:90 suppresses
  16. L14:140 suppresses
  17. L20:433 suppresses
  18. L23:262 suppresses
  19. L16:1421 suppresses
  20. L20:499 excites
  21. L24:141 suppresses
  22. L23:864 suppresses
  23. L24:94 excites
  24. L25:17 excites

Feeds later layers

  1. L27:921 suppresses
  2. L27:135 weakly suppresses
  3. L27:527 weakly suppresses
  4. L27:802 weakly suppresses
  5. L27:2890 weakly excites
  6. L27:580 weakly excites
  7. L27:526 weakly suppresses
  8. L27:476 weakly suppresses
  9. L27:514 weakly suppresses
  10. L27:912 weakly suppresses
  11. L27:553 weakly suppresses
  12. L27:898 weakly excites
  13. L27:2035 weakly excites
  14. L27:186 weakly suppresses
  15. L27:450 weakly excites
  16. L27:573 weakly suppresses
  17. L27:213 weakly excites
  18. L27:337 weakly suppresses
  19. L27:327 weakly excites
  20. L27:722 weakly suppresses
  21. L27:485 weakly suppresses
  22. L27:561 weakly suppresses
  23. L27:1982 weakly excites
  24. L27:458 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 -19.75 and -18.523 values between -13.60 and -12.371 values between -12.37 and -11.133 values between -11.13 and -9.903 values between -9.90 and -8.674 values between -8.67 and -7.449 values between -7.44 and -6.2111 values between -6.21 and -4.989 values between -4.98 and -3.7519 values between -3.75 and -2.5235 values between -2.52 and -1.291687 values between -1.29 and -0.062041 values between -0.06 and 1.1830 values between 1.18 and 2.4117 values between 2.41 and 3.647 values between 3.64 and 4.8710 values between 4.87 and 6.105 values between 6.10 and 7.333 values between 7.33 and 8.561 values between 8.56 and 9.791 values between 9.79 and 11.021 values between 12.25 and 13.481 values between 13.48 and 14.711 values between 14.71 and 15.951 values between 15.95 and 17.182 values between 17.18 and 18.411 values between 27.02 and 28.251 values between 28.25 and 29.48offnegativepositiveoften

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. 촌
  2. 一期
  3. 两岸
  4. tile
  5. 化合
  6. 嫉
  7. -feature
  8. generation
  9. donn
  10. 省政府

Suppresses

  1. ing
  2. ging
  3. rys
  4. ming
  5. ries
  6. ister
  7. azer
  8. ges
  9. nelle
  10. медицинск