Neuron Explorer

Qwen3-0.6B

'other' in 'in other words'

A neuron in layer 22 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 'other' when it comes right after 'in' and leads into a restatement, as in 'in other words'. Turning it up made the model repeat the word 'words' over and over.

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

    carriers showed significantly worse prospective memories. Far more often than non-carriers, they failed to remember to write down the desired word when they were supposed to – in other words, they forgot to do what they meant to do, when they meant to

  2. #2 peak

    code that can be hosted by the CLR, while at the same time can be accessed in a uniform manner by all languages that the .NET platform support. In other words, CLS is a subset of the full functionality of CTS.⏎-

  3. #3 peak

    - Compassion vs Empathy: What Is the Difference?⏎What Is Empathy?⏎Empathy is the ability to put yourself in someone else’s shoes. In other words, it’s the ability to feel or understand another person. Experts suggest it

  4. #4 peak

    a number of signals with rather close frequencies. As soon as one of these signals "exceeds" the others, it masks them more or less: in other words, the other sound are perceived as much weaker than they actually are, or

  5. #5 peak

    “spiritual aristocrats” Tetraphilosophy appear simple⏎with capacious dialectic models and abbreviations that are clear to any young person; in other words, we have democracy instead of aristocr

  6. #6 peak

    have various causes, ranging in severity from 'worrying' to 'generally fatal'. Finding the true cause means ruling out or confirming each possibility – in other words, diagnosis.⏎Diagnosis is usually a complex process due to the sheer number

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 first and the second antumnumus. I mean, the first antumnumus is the antumnus antumnumus antumnumus... that antumnumum antumnum

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

    submit original poems. The only constraint is that the poems must follow a format suitable for Twitter–280 characters or fewer. These bite-sized verses often prove interesting, complex,

  2. #2 peak

    online schedule and for those that don’t have the ability to move to another state, the college they wish to attend is just a few mouse clicks away.⏎Is this where grade-school level education is headed? I can’t say for certain that is

  3. #3 peak

    advance through your welding career, you will need to do various complicated welding jobs. So, learning how to weld pipes is necessary to avoid any mistakes or accidents.⏎This job can become even more challenging when it is welding pipes instead of flat surfaces.

  4. #4 peak

    .⏎This decision came down from the top brass. Don't bother questioning it.⏎I would not use it to talk about the people who administer a program.⏎Is this use of "brass" common in the UK?⏎I'm not a teacher

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. L20:1065 excites
  2. L19:1587 suppresses
  3. L19:2150 excites
  4. L11:1687 suppresses
  5. L21:1463 suppresses
  6. L6:2787 weakly suppresses
  7. L21:2820 weakly excites
  8. L21:708 weakly suppresses
  9. L19:1926 weakly suppresses
  10. L18:1537 weakly suppresses
  11. L9:2958 weakly excites
  12. L17:576 weakly excites
  13. L13:855 weakly excites
  14. L3:2393 weakly excites
  15. L3:993 weakly suppresses
  16. L7:2134 weakly suppresses
  17. L12:148 weakly excites
  18. L16:973 weakly suppresses
  19. L16:3002 weakly excites
  20. L5:86 weakly suppresses
  21. L1:2581 weakly excites
  22. L14:1692 weakly suppresses
  23. L20:1577 weakly suppresses
  24. L21:998 weakly suppresses

Feeds later layers

  1. L27:1957 strongly excites
  2. L23:2317 excites
  3. L26:523 excites
  4. L23:2612 excites
  5. L26:556 excites
  6. L25:2535 excites
  7. L25:1331 weakly excites
  8. L24:268 weakly excites
  9. L23:1471 weakly excites
  10. L26:2130 weakly excites
  11. L25:991 weakly excites
  12. L23:2546 weakly excites
  13. L25:1822 weakly excites
  14. L25:2866 weakly excites
  15. L24:1841 weakly excites
  16. L24:2465 weakly excites
  17. L24:703 weakly excites
  18. L23:2872 weakly excites
  19. L24:1594 weakly excites
  20. L23:2588 weakly excites
  21. L27:1667 weakly excites
  22. L25:1571 weakly excites
  23. L23:409 weakly excites article 'a' after a role or title
  24. L24:1500 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 -7.15 and -6.452 values between -6.45 and -5.741 values between -5.04 and -4.334 values between -4.33 and -3.635 values between -3.63 and -2.927 values between -2.92 and -2.2216 values between -2.22 and -1.5174 values between -1.51 and -0.811272 values between -0.81 and -0.102311 values between -0.10 and 0.60148 values between 0.60 and 1.3121 values between 1.31 and 2.0111 values between 2.01 and 2.7210 values between 2.72 and 3.428 values between 3.42 and 4.135 values between 4.13 and 4.832 values between 4.83 and 5.543 values between 5.54 and 6.241 values between 6.24 and 6.952 values between 7.65 and 8.361 values between 8.36 and 9.061 values between 9.06 and 9.771 values between 12.59 and 13.291 values between 20.34 and 21.05offnegativepositiveoften

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. words
  2. words
  3. Words
  4. Words
  5. word
  6. _words
  7. Word
  8. word
  9. .words
  10. Word

Suppresses

  1. 硌
  2. MaxY
  3. INESS
  4. asant
  5. directory
  6. ragment
  7. major
  8. ıs
  9. glGet
  10. ighth