Neuron Explorer

Qwen3-0.6B

the 'pre' in pre-words

A neuron in layer 26 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 start 'pre' of words like prehistoric, prehistory and pre-Civil War, where the word is split and 'pre' comes first. When it was turned up, the text drifted toward unrelated fragments and repeated phrases, so the push did not produce a clear theme.

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

    States.⏎Kahlo’s artistic sensibility and signature style developed in full. Her main source of inspiration was Mexican folk culture, which she blended with elements from pre-Colombian and Catholic mythology in her paintings.<|endoftext|>|Wednesday, 3

  2. #2 peak

    apparent spatial dimension to the site, it appeared possible that activity specific areas might be detected. Other reasons for selecting this site included the absence of knowledge concerning the late prehistory of this area, the rapid rate of erosion and destruction occuring

  3. #3 peak

    2011 - Nightlife during the Mesozoic era was full of activity and fraught with dangers, suggests a new study that found some dinosaurs and prehistoric reptiles were nocturnal, with carnivorous dinosaurs likely sneaking up on

  4. #4 peak

    while the old countries, populated since prehistoric times, each have their great groups of related motifs: the supernatural light creature

  5. #5 peak

    describes, “the piers of the bridge were festooned with statuary of mermen on dolphins. The bridge balustrades were ornately molded in pre-cast faux stone and lighting was generously provided from electric globes on elebor

  6. #6 peak

    It’s a 50-episode story that follows the time-traveling adventures of a little boy and his trusty computer companion, as they explore the prehistoric world. Listen to the episode here.⏎Now that you’ve listened to the

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: a one a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a

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

    to the very essence of how the Universe expresses itself.⏎He begins by stating that all tathagatas are without Buddha nature although they have arrived in the state of reality. A

  2. #2 peak

    works such as Prayers or Meditations in 1545, she became not only the first English Queen

  3. #3 peak

    . People are strongly advised to take special care if a water supply contains two VOCs, however, even if they are both below the maximum allowable concentration.⏎Bioworx has just come out<|endoftext|>Nikola Tesla Articles⏎Tesla, a

  4. #4 peak

    Nature Beyond Conditions and a commentary by Rev. David Xi-Ken Astor, Sensei⏎BUDDHA NATURE BEYOND CONDITIONS⏎“All tathagatas are without Buddha nature, but at the same time, previously they have fully

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. L16:25 weakly excites
  2. L2:1529 weakly excites
  3. L1:2340 weakly excites
  4. L23:206 weakly excites
  5. L8:10 weakly suppresses
  6. L16:110 weakly suppresses
  7. L17:1597 weakly suppresses
  8. L21:463 weakly suppresses
  9. L20:197 weakly suppresses
  10. L5:139 weakly excites
  11. L8:1802 weakly excites
  12. L18:2052 weakly suppresses
  13. L16:71 weakly excites
  14. L17:162 weakly suppresses
  15. L16:36 weakly suppresses
  16. L20:728 weakly excites
  17. L4:731 weakly suppresses
  18. L18:45 weakly suppresses
  19. L6:2269 weakly suppresses
  20. L10:1155 weakly suppresses
  21. L19:2685 weakly excites
  22. L8:3 weakly suppresses
  23. L6:468 weakly excites
  24. L16:15 weakly suppresses

Feeds later layers

  1. L27:797 suppresses
  2. L27:709 weakly suppresses
  3. L27:663 weakly suppresses
  4. L27:231 weakly excites
  5. L27:2732 weakly suppresses
  6. L27:2917 weakly excites
  7. L27:2474 weakly excites
  8. L27:2819 weakly suppresses
  9. L27:2029 weakly suppresses
  10. L27:1754 weakly excites
  11. L27:1143 weakly excites
  12. L27:644 weakly excites
  13. L27:1435 weakly excites
  14. L27:1472 weakly excites
  15. L27:1296 weakly suppresses
  16. L27:2883 weakly excites
  17. L27:1716 weakly excites
  18. L27:2003 weakly suppresses
  19. L27:1737 weakly suppresses
  20. L27:755 weakly suppresses
  21. L27:2587 weakly suppresses
  22. L27:1037 weakly excites
  23. L27:567 weakly excites
  24. L27:2236 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 -18.27 and -17.381 values between -15.60 and -14.711 values between -14.71 and -13.821 values between -13.82 and -12.943 values between -12.94 and -12.055 values between -12.05 and -11.166 values between -11.16 and -10.277 values between -10.27 and -9.3810 values between -9.38 and -8.5012 values between -8.50 and -7.6120 values between -7.61 and -6.7221 values between -6.72 and -5.8343 values between -5.83 and -4.9464 values between -4.94 and -4.0581 values between -4.05 and -3.17109 values between -3.17 and -2.28179 values between -2.28 and -1.39327 values between -1.39 and -0.501370 values between -0.50 and 0.39582 values between 0.39 and 1.28338 values between 1.28 and 2.17206 values between 2.17 and 3.05168 values between 3.05 and 3.94107 values between 3.94 and 4.8377 values between 4.83 and 5.7258 values between 5.72 and 6.6136 values between 6.61 and 7.5028 values between 7.50 and 8.3815 values between 8.38 and 9.2712 values between 9.27 and 10.169 values between 10.16 and 11.054 values between 11.05 and 11.941 values between 11.94 and 12.822 values between 13.71 and 14.602 values between 14.60 and 15.492 values between 16.38 and 17.27offnegativepositiveoften

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. erate
  2. 虮
  3. состояние
  4. 骊
  5. Lud
  6. 期待
  7. 啡
  8. 工业大学
  9. �
  10. atif

Suppresses

  1. 一段
  2. 一个小
  3. 一个
  4. a
  5. 一个人
  6. ache
  7. 一個
  8. 的一个
  9. 一个多
  10. 一份