Neuron Explorer

Qwen3-0.6B

the word 'or' between alternatives

A neuron in layer 15 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 lights up on the word 'or' when it offers a second option, like 'a piece of white cardstock or poster board'. Turning it up did not produce a clear theme in the generated text.

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

    Dane breed comes from German descent without any known associations to the country of Denmark, at least that’s what canine historians say.⏎The breed might be four centuries old or even more, but the actual origins of the breed are murky. Great Danes

  2. #2 peak

    ; ribs 5, acute or round-obtuse; vittae 1 in each furrow, 2 on commissure. Seed face plane or slightly concave. Carpoph

  3. #3 peak

    edges of hedgerows and woodland, and by the roadside. Wild chervil blooms from mid spring to early summer. It is an herbaceous biennial or short-lived perennial plant in the family Apiaceae (Umbellifera

  4. #4 peak

    2012.⏎Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.3 or any later version published by the Free Software Foundation; with no Invariant Sections,

  5. #5 peak

    a life-like silhouette.⏎In a darkened room, have an individual stand sideways against a wall. Behind⏎the person, tape a piece of white cardstock or poster board on the wall.⏎The board should start about mid-chest and go

  6. #6 peak

    a record of a David de Boivil witnessing a charter as early as 1164.<|endoftext|>For the questions below, use the most efficient Excel formula or procedure to find your answer. Calculations by hand (or calculator) should only

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 same set of algorithms are being used to calculate the same set of data in the same set of set of data set, which the same set of data set, the same set of set of data set, the same set of set of

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

    murex shellfish, known as Tyrian purple. This color was, in many cultures of ancient times, reserved for the use of royalty, or at least nobility.⏎It was often attacked by Egypt, besieged by Shalman

  2. #2 peak

    appear: cataract and AMD (Age-Related Macular Degeneration) are the most limiting. But it is possible to prevent them or, at least, to

  3. #3 peak

    Cinderella a folktale or fairy tale?⏎Cinderella is considered a fairy tale, not a folk tale. Folk tales have a basis in reality or something which happened in history.⏎What is a common event in a folktale<|endoftext|>

  4. #4 peak

    voted down.⏎Jay founded the New-York Society for Promoting the Manumission of Slaves, and Protecting Such of Them as Have Been, or May be Liberated or the New York Manumission Society, and became its first

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. L10:135 strongly excites 'or' in alternative requirements
  2. L4:108 excites "or" offering a choice in web text
  3. L11:273 suppresses 'and' joining clauses in instructions
  4. L7:1475 excites 'or' joining alternatives in lists
  5. L13:1208 excites
  6. L14:2065 excites
  7. L13:553 suppresses
  8. L12:836 suppresses
  9. L10:901 suppresses
  10. L1:1710 suppresses
  11. L3:1200 excites
  12. L13:411 excites
  13. L7:1646 suppresses
  14. L11:406 suppresses
  15. L14:2014 excites
  16. L12:1094 excites
  17. L11:1861 weakly suppresses
  18. L6:733 weakly suppresses
  19. L9:2010 weakly suppresses
  20. L12:1452 weakly suppresses
  21. L12:2393 weakly excites
  22. L0:198 weakly excites
  23. L10:891 weakly suppresses
  24. L9:352 weakly suppresses

Feeds later layers

  1. L16:925 excites
  2. L18:1138 excites
  3. L17:2166 excites
  4. L16:1782 excites
  5. L16:935 suppresses
  6. L19:1343 excites
  7. L16:1959 excites
  8. L16:2387 suppresses
  9. L16:1913 suppresses
  10. L16:1068 suppresses
  11. L16:2289 excites 'or' linking two alternatives
  12. L16:1181 weakly excites
  13. L16:1735 weakly excites
  14. L21:1974 weakly excites
  15. L16:2818 weakly suppresses
  16. L19:2597 weakly excites
  17. L16:420 weakly excites
  18. L22:266 weakly excites
  19. L16:678 weakly suppresses
  20. L18:2197 weakly excites
  21. L24:1184 weakly excites
  22. L16:2261 weakly excites
  23. L16:1492 weakly suppresses
  24. L20:2429 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 -3.79 and -3.442 values between -2.37 and -2.026 values between -2.02 and -1.666 values between -1.66 and -1.3114 values between -1.31 and -0.9548 values between -0.95 and -0.60308 values between -0.60 and -0.242757 values between -0.24 and 0.12686 values between 0.12 and 0.4745 values between 0.47 and 0.836 values between 0.83 and 1.186 values between 1.18 and 1.541 values between 1.54 and 1.894 values between 1.89 and 2.252 values between 2.25 and 2.604 values between 2.60 and 2.961 values between 3.31 and 3.671 values between 3.67 and 4.022 values between 4.38 and 4.731 values between 6.86 and 7.221 values between 7.57 and 7.933 values between 7.93 and 8.281 values between 8.64 and 8.992 values between 10.06 and 10.41offnegativepositiveoften

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. 까
  2. other
  3. ichte
  4. 别的
  5. else
  6. simply
  7. /dd
  8. 仅仅
  9. IDs
  10. ENDING

Suppresses

  1. seriousness
  2. GridBagConstraints
  3. ynchron
  4. occupies
  5. 有大量的
  6. -Length
  7. ї
  8. Loves
  9. 走下去
  10. -'+