Neuron Explorer

Qwen3-0.6B

unclear: 'script' and boilerplate links

A neuron in layer 1 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 weakly fires on the word 'script' after 'This' in academic writing and on 'PDF' in link labels. It dips on website navigation text such as 'Skip to main content'. Turning it up or down gave writing with no obvious link to these patterns.

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

    Asia are, of the Brahmi script. This script was developed through various stages. It is closely related to northern Nagari which took

  2. #2 peak

    ''s most intelligent animals. See a great silverback gorilla, the monkey with

  3. #3 peak

    to facilitate children’s science exploration through everyday kitchen substances and experiences such as mixing, pouring, heating, and cooling. Full text» [PDF, 162KB]⏎Marvelous Explorations Through Science and Stories

  4. #4 peak

    hakusai and hyakusai 7 of the two Chinese characters forming it, has been the object of various interpretations, mainly by Oriental scholars.⏎This script does not aim to solve the problem in an irrefutable way, but simply

  5. #5 peak

    (2) failing to provide food, care or shelter;⏎(3) abandoning an animal

  6. #6 peak

    to make some furtive words more readable. Also the spacing, boldness and capital case can be switched to make the letters stand out from their context. Optionally a handy hover-cross is shown, highlighting those letters that are in line with the

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 user of this question is the one that is not provided directly. Give me a answer. The question is here. Here, the user of this question is the one that is not provided directly. Give me a answer. This question

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

    same object.⏎Random House Unabridged Dictionary, Copyright © 1997, by Random House, Inc., on Infoplease.<|endoftext|>Skip to Content⏎Nodding thistle is one of the most competitive thistle species

  2. #2 peak

    percentages and probability as a tool in materializing ideas (game mechanics)⏎4. Playtesting for balancing the game (game mechanics/ dynamics)<|endoftext|>Skip to main content⏎More Search Options⏎A member of our team will call you back within one

  3. #3 peak

    offer support, and learn more about the incident. If needed, the Dean’s team will direct the report to the appropriate office for investigation.<|endoftext|>Skip to main content⏎Create interactive lessons using any digital content including wikis with our free sister product

  4. #4 peak

    to supply either plain text or an HTML representation. In the example, the client prefers HTML but is willing to accept plain text.⏎Accept: text/html, text/plain⏎Accept-Charset: Specifies the character set,

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. L0:1301 suppresses
  2. L0:1224 suppresses
  3. L0:178 suppresses
  4. L0:592 weakly excites
  5. L0:16 weakly excites
  6. L0:706 weakly suppresses
  7. L0:1 weakly suppresses
  8. L0:2254 weakly suppresses
  9. L0:2187 weakly suppresses
  10. L0:711 weakly excites
  11. L0:1092 weakly excites
  12. L0:640 weakly suppresses abbreviation periods after initials
  13. L0:2562 weakly excites
  14. L0:2973 weakly suppresses
  15. L0:508 weakly suppresses
  16. L0:1184 weakly excites
  17. L0:85 weakly suppresses
  18. L0:2151 weakly suppresses
  19. L0:491 weakly excites
  20. L0:1993 weakly excites
  21. L0:213 weakly excites
  22. L0:4 weakly suppresses
  23. L0:953 weakly suppresses
  24. L0:2940 weakly excites

Feeds later layers

  1. L2:1571 suppresses
  2. L23:2580 weakly suppresses
  3. L20:826 weakly suppresses
  4. L5:780 weakly suppresses
  5. L5:1442 weakly suppresses
  6. L22:1841 weakly excites
  7. L14:3052 weakly suppresses
  8. L21:1803 weakly suppresses
  9. L3:2157 weakly suppresses
  10. L25:1216 weakly suppresses
  11. L14:1090 weakly suppresses
  12. L6:560 weakly excites
  13. L20:2163 weakly suppresses
  14. L14:2978 weakly excites
  15. L7:1092 weakly excites
  16. L21:1570 weakly suppresses
  17. L6:1548 weakly suppresses
  18. L7:1855 weakly suppresses
  19. L6:1902 weakly suppresses
  20. L18:291 weakly suppresses
  21. L21:2440 weakly suppresses
  22. L22:1888 weakly suppresses
  23. L9:3011 weakly suppresses
  24. L9:1536 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 -0.38 and -0.372 values between -0.29 and -0.275 values between -0.27 and -0.263 values between -0.26 and -0.241 values between -0.24 and -0.223 values between -0.20 and -0.188 values between -0.18 and -0.1616 values between -0.16 and -0.1431 values between -0.14 and -0.1335 values between -0.13 and -0.1144 values between -0.11 and -0.0957 values between -0.09 and -0.07143 values between -0.07 and -0.05285 values between -0.05 and -0.03616 values between -0.03 and -0.011393 values between -0.01 and 0.01602 values between 0.01 and 0.02255 values between 0.02 and 0.04146 values between 0.04 and 0.0686 values between 0.06 and 0.0857 values between 0.08 and 0.1030 values between 0.10 and 0.1232 values between 0.12 and 0.1418 values between 0.14 and 0.158 values between 0.15 and 0.175 values between 0.17 and 0.199 values between 0.19 and 0.218 values between 0.21 and 0.235 values between 0.23 and 0.251 values between 0.25 and 0.261 values between 0.28 and 0.301 values between 0.30 and 0.321 values between 0.34 and 0.36offnegativepositiveoften

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. est
  2. ffer
  3. igu
  4. stral
  5. ist
  6. Filter
  7. cret
  8. str
  9. A

Suppresses

  1. !(:
  2. MBER
  3. spacing
  4. uptools
  5. Podesta
  6. Brigham
  7. ことに
  8. ENCE
  9. 意义上
  10. !.⏎⏎