Neuron Explorer

Qwen3-0.6B

'Me' starting a surname

A neuron in layer 25 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 the syllable 'Me' when it starts a name, such as Meier, Meade or Meinkoth, mostly surnames inside news and citation text. Steering it changed the topic a lot without a clear link to names, so the link to this trigger is only a guess.

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

    Want to Be President? is chock-full of interesting tidbits about our Presidents;...⏎Activities and Games, Extension Activities, Book Resources⏎Extension activity for Arthur Meets the President by Marc Brown.⏎Biography of former United States President George Herbert

  2. #2 peak

    described are explored⏎Footnote 1: adapted from Woodward-Kron, R., Thomson, E. & Meek, J. (2000) A text based guide to academic

  3. #3 peak

    called poults. A group of turkeys is called a flock.⏎The benefits of raising white turkeys are<|endoftext|>In this thoroughly researched history, Graydon Meints tells the fascinating story of the railroad’s arrival and development in Michigan. An

  4. #4 peak

    of his campaign, he had to protect 150 miles of<|endoftext|>Rethinking Teen Sex⏎By Tim Brady⏎A new study by assistant professor Ann Meier has made waves with its fresh look at adolescent sexuality⏎Almost half of American

  5. #5 peak

    (Charonia Research, 1996b; Charonia Research, 1996c; Hendler, 1995; Meinkoth, 1981)⏎This is a bottom dwelling species that

  6. #6 peak

    , it burned only the energy that was available from the environment, slowing down the ‘fire of life’ in times when resources became scarce,” said study co-author Meike Kohler, a paleobiologist at the Autonomous University of Barcelona, Spain

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

    timing circuit that gives precise and highly stable delays of time or oscillation. These types of ICs are very cheap and reliable in cost when we compared with the OP-Amp applications in the same areas. These ICs are used as an ast

  2. #2 peak

    next step is to determine the identifier for the drive you want to make an image of. In the terminal window, type in the command⏎sudo fdisk -l⏎Our 1GB drive is /<|endoftext|>During the period from 19

  3. #3 peak

    use any editor such as edit to write a C program.⏎C:> edit myprogram.c⏎- After written the program, to compile it, issue the command cl⏎C:> cl myprogram.c⏎- During the compilation and linking, my

  4. #4 peak

    agonists (Allbuterol/Salbutamol, Proventil, Ventolin) and theophylline

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. L20:1257 excites
  2. L20:831 excites
  3. L13:44 suppresses inside web addresses and citations
  4. L10:155 excites
  5. L24:214 excites
  6. L15:184 excites
  7. L20:2046 excites
  8. L21:720 excites
  9. L12:158 suppresses
  10. L20:54 excites
  11. L11:223 excites
  12. L21:2711 excites
  13. L19:1685 suppresses
  14. L20:162 excites
  15. L18:251 excites
  16. L23:1936 suppresses
  17. L23:126 suppresses
  18. L19:400 excites
  19. L8:328 excites
  20. L13:2366 suppresses
  21. L10:669 suppresses
  22. L18:1292 excites
  23. L21:2617 suppresses
  24. L9:74 excites

Feeds later layers

  1. L27:2277 strongly suppresses
  2. L27:1606 suppresses
  3. L27:1050 suppresses
  4. L27:2506 suppresses
  5. L26:408 weakly suppresses
  6. L26:440 weakly suppresses
  7. L26:2147 weakly suppresses
  8. L27:2055 weakly excites
  9. L27:1229 weakly excites
  10. L26:706 weakly excites
  11. L26:1474 weakly suppresses
  12. L27:1917 weakly excites
  13. L27:597 weakly excites
  14. L27:1488 weakly suppresses
  15. L26:2671 weakly excites
  16. L26:1369 weakly excites 'Pr' starting capitalised names
  17. L26:2641 weakly suppresses
  18. L26:2255 weakly suppresses
  19. L26:506 weakly excites
  20. L26:2998 weakly excites
  21. L27:908 weakly excites
  22. L27:3034 weakly excites
  23. L26:246 weakly suppresses
  24. L27:2559 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 -2.78 and -2.341 values between -2.34 and -1.891 values between -1.89 and -1.4512 values between -1.45 and -1.0074 values between -1.00 and -0.561343 values between -0.56 and -0.112289 values between -0.11 and 0.34170 values between 0.34 and 0.7811 values between 0.78 and 1.231 values between 1.23 and 1.671 values between 1.67 and 2.122 values between 2.12 and 2.561 values between 4.34 and 4.791 values between 14.58 and 15.02offnegativepositiveoften

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. edException
  2. etail
  3. ritis
  4. passwd
  5. ByExample
  6. ponge
  7. onSelect
  8. unky
  9. ting
  10. igit

Suppresses

  1. 来
  2. 來
  3. 去
  4. 来做
  5. come
  6. comes
  7. 来的
  8. 来找
  9. coming
  10. 来进行