Neuron Explorer

Qwen3-0.6B

'be' after 'to' or 'need to'

A neuron in layer 20 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 'be' when it comes after 'to' or 'need to', as in 'have time to be' or 'expects to be ready'. Turning it up made the model write in a loose, repetitive way, so the steering results don't clearly show what it does.

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

    ⏎Why Lesson Planet?⏎Find quality lesson planning resources, fast!⏎Share & remix collections to collaborate.⏎Organize your curriculum with collections. Easy!⏎Have time to be more creative & energetic with your students!⏎Students are introduced to the characteristics of a

  2. #2 peak

    ⏎Why Lesson Planet?⏎Find quality lesson planning resources, fast!⏎Share & remix collections to collaborate.⏎Organize your curriculum with collections. Easy!⏎Have time to be more creative & energetic with your students!⏎Students explore the Orange Revolution. In this

  3. #3 peak

    Why Lesson Planet?⏎Find quality lesson planning resources, fast!⏎Share & remix collections to collaborate.⏎Organize your curriculum with collections. Easy!⏎Have time to be more creative & energetic with your students!⏎Students participate in a variety of physical activities

  4. #4 peak

    <|endoftext|>Why Lesson Planet?⏎Find quality lesson planning resources, fast!⏎Share & remix collections to collaborate.⏎Organize your curriculum with collections. Easy!⏎Have time to be more creative & energetic with your students!⏎In this verb tense review worksheet, students

  5. #5 peak

    a Dragon spacecraft from the Cape Canaveral Air Force Station. The successful flight demonstrated the components necessary to transport humans to low-Earth orbit. SpaceX expects to be ready to fly its first manned mission in 2014.⏎Dragon is

  6. #6 peak

    -10. The full text can be found on the Board of Studies website.⏎Students must study at least 100 hours in one language, to be completed over one continuous 12-month period. When students have completed the mandatory

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: I do it, it means it does it, it does it, it means it does it. I mean it does it it *does* it it *does* it it *does* it. The it works it *does*

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

    food and shelter. Fruits such as acorns or berries are vitally important, but sometimes other benefits are less noticeable. Many animals are host dependent--that is, they need a specific type of plant to survive. Butterfly larvae may chew on

  2. #2 peak

    . Recent research by Dr. Joseph Allen, a professor at the University of Virginia, points out the benefits of certain types of arguing between parents and children. That is, some forms of arguing actually strengthen

  3. #3 peak

    the Sensee River. Some of the men escaped the hail of bullets, but few⏎of the horses escaped. Tender-hearted gunners took time out to<|endoftext|>Is PET Plastic Safe for Food and Beverage?⏎Is PET plastic safe?

  4. #4 peak

    your personal desktop machine or laptop, and to files located on a remote web hosting server. We know how important it is to keep all the data safe. That is why at NTC Hosting we have developed a powerful backup solution, which takes care

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. L18:960 strongly suppresses
  2. L17:227 strongly suppresses
  3. L19:1115 excites
  4. L19:576 suppresses
  5. L4:20 suppresses
  6. L15:599 excites
  7. L9:66 excites
  8. L16:466 excites 'is' after a long subject phrase
  9. L4:2128 suppresses
  10. L19:1612 excites
  11. L3:1301 excites
  12. L10:1736 suppresses
  13. L16:2038 suppresses
  14. L0:207 excites
  15. L18:612 weakly suppresses
  16. L17:2353 weakly excites
  17. L17:2050 weakly excites the word 'be' after modals
  18. L18:1167 weakly suppresses
  19. L6:2237 weakly excites
  20. L4:117 weakly excites
  21. L13:775 weakly suppresses
  22. L18:144 weakly suppresses
  23. L13:612 weakly suppresses
  24. L18:391 weakly excites

Feeds later layers

  1. L22:550 strongly excites
  2. L27:1368 excites
  3. L23:360 suppresses
  4. L21:1555 excites
  5. L22:2991 excites
  6. L23:576 excites
  7. L23:1160 excites
  8. L21:1814 excites
  9. L22:415 suppresses
  10. L21:1872 excites
  11. L21:1434 excites
  12. L26:395 excites
  13. L23:473 excites
  14. L23:1293 excites
  15. L21:1047 excites
  16. L23:1083 excites
  17. L23:1004 excites
  18. L26:571 excites
  19. L27:217 excites
  20. L25:2151 excites
  21. L23:628 excites
  22. L21:1675 excites
  23. L26:237 excites
  24. L23:1012 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 -15.10 and -13.301 values between -6.09 and -4.29498 values between -2.49 and -0.693270 values between -0.69 and 1.1116 values between 1.11 and 2.927 values between 2.92 and 4.724 values between 4.72 and 6.522 values between 6.52 and 8.325 values between 8.32 and 10.128 values between 10.12 and 11.937 values between 11.93 and 13.736 values between 13.73 and 15.534 values between 15.53 and 17.337 values between 17.33 and 19.133 values between 19.13 and 20.9311 values between 20.93 and 22.7410 values between 22.74 and 24.543 values between 24.54 and 26.346 values between 26.34 and 28.1410 values between 28.14 and 29.944 values between 29.94 and 31.744 values between 31.74 and 33.553 values between 33.55 and 35.354 values between 35.35 and 37.152 values between 37.15 and 38.954 values between 38.95 and 40.751 values between 40.75 and 42.563 values between 42.56 and 44.361 values between 44.36 and 46.161 values between 53.37 and 55.172 values between 55.17 and 56.97offnegativepositiveoften

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. considered
  2. flown
  3. able
  4. ready
  5. taken
  6. cie
  7. gone
  8. worth
  9. absorbed
  10. supposed

Suppresses

  1. edList
  2. must
  3. 日消息
  4. befind
  5. ющую
  6. 眼镜
  7. rop
  8. drop
  9. очередь
  10. 一点