Neuron Explorer

Qwen3-0.6B

'the more X' clause before 'the more'

A neuron in layer 19 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 reacts to the comma that ends a phrase like 'the more you drink, the more...', where the model is about to write the matching 'the more' half. When it was turned up, the model fell into a loop of 'the more the more the better'.

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

    this is done, the company loses vital manufacturing and quality information. This information accurately collected is the only basis of implementing quality improvement. The more granular the information, the more effective it can become.⏎If the only options are in-tolerance and

  2. #2 peak

    are being used to validate this identity combination, and to prove a virtual person is indeed the same person in real life. The more value an identity has access to, more effort and work is required to validate and establish secure authentications for it.⏎

  3. #3 peak

    2 studies, found there was a dose-response association between intake of plant protein and risk of death (from any cause) - the more plant protein in the diet, the lower the risk of death.⏎What foods are high in plant protein? Leg

  4. #4 peak

    the file names. ACTOZ is the name we shall use throughout this user's manual.⏎Actoz is a very powerful instrument. The more data it contains, the greater its potential for thorough analysis and unexpected results. One of the greatest qualities

  5. #5 peak

    few (or 30) seconds or your water’s a couple degrees off temperature, you’re NOT going to ruin your film. The more accurate you are,

  6. #6 peak

    stress.⏎Stress is an underlying factor in illness as it reduces your immune system’s ability to fight off antigens. This means that the more stressed you are, the likelier you are to pick up infections and fight illness.⏎Therefore, a

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: If I can get back to old friends in 2020, then maybe it would help to keep working at school now, by working as much as possible, but instead of working for money, working to save money in working.

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

    a fraction of a second.⏎The discovery has been accepted for publication in Physical Review Letters.⏎Gravitational waves are ripples in the universe caused by some of the most energetic cosmic events, from exploding stars to supermassive black hole mergers

  2. #2 peak

    two end systems⏎B. It specifies the quality of the data & connection⏎C. It specify the size of data⏎D. All of the above⏎6-4. The class-based addressing is also known as⏎A

  3. #3 peak

    a unique continental and leading agricultural product. One of the countries on the equator line benefits from its climate and agriculture production. Before you

  4. #4 peak

    research.⏎LaMantia and his colleagues looked at brain tissue from mice missing the same region. They found that the mice have unusually few neurons in two of the six layers of the cerebral cortex — the outer portion of the brain. These neurons

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. L18:2316 excites
  2. L17:30 suppresses
  3. L4:52 excites
  4. L16:2248 suppresses
  5. L18:2979 excites
  6. L9:24 excites "the" before a named organisation or title
  7. L16:419 excites
  8. L18:641 suppresses
  9. L12:435 suppresses
  10. L15:1848 weakly suppresses
  11. L18:2308 weakly suppresses
  12. L12:3007 weakly suppresses
  13. L2:1473 weakly excites
  14. L18:239 weakly suppresses
  15. L11:1168 weakly excites
  16. L8:636 weakly suppresses
  17. L5:125 weakly excites "of" after all three, both, ten
  18. L6:426 weakly suppresses
  19. L18:602 weakly suppresses
  20. L4:73 weakly suppresses
  21. L4:2719 weakly suppresses
  22. L10:1377 weakly suppresses
  23. L7:63 weakly excites
  24. L3:59 weakly excites

Feeds later layers

  1. L26:507 strongly excites
  2. L27:127 strongly suppresses
  3. L27:399 strongly excites
  4. L22:171 strongly excites
  5. L24:49 strongly suppresses
  6. L21:863 excites
  7. L21:310 excites
  8. L21:2949 excites comma closing 'the more X, the more Y' clauses
  9. L27:121 suppresses
  10. L20:64 suppresses
  11. L21:758 excites
  12. L25:247 excites
  13. L22:197 suppresses
  14. L22:353 suppresses
  15. L23:94 suppresses
  16. L20:593 excites
  17. L20:2278 excites
  18. L20:738 excites
  19. L24:92 suppresses
  20. L20:2321 excites
  21. L25:1796 excites
  22. L22:360 weakly excites
  23. L26:693 weakly excites
  24. L22:1589 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 -14.52 and -13.471 values between -12.43 and -11.391 values between -11.39 and -10.341 values between -10.34 and -9.302 values between -9.30 and -8.264 values between -8.26 and -7.224 values between -7.22 and -6.175 values between -6.17 and -5.1318 values between -5.13 and -4.0932 values between -4.09 and -3.0433 values between -3.04 and -2.0062 values between -2.00 and -0.961922 values between -0.96 and 0.091754 values between 0.09 and 1.1322 values between 1.13 and 2.1713 values between 2.17 and 3.225 values between 3.22 and 4.264 values between 4.26 and 5.308 values between 5.30 and 6.346 values between 6.34 and 7.392 values between 7.39 and 8.432 values between 8.43 and 9.472 values between 10.52 and 11.561 values between 11.56 and 12.601 values between 13.65 and 14.691 values between 14.69 and 15.731 values between 26.16 and 27.20offnegativepositiveoften

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. the
  2. the
  3. →the
  4. _the
  5. The
  6. .the
  7. The
  8. THE
  9. thed
  10. …the

Suppresses

  1. RIEND
  2. awning
  3. REEN
  4. ��
  5. ày
  6. reen
  7. 余人
  8. Bosch
  9. emics
  10. uplicates