Neuron Explorer

Qwen3-0.6B

'of' in formal noun phrases, mixed

A neuron in layer 12 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 word 'of' that follows a noun in formal writing, such as 'a substance of one density' or 'warranties of any kind'. It is pushed the other way by fixed phrases like 'Conflicts of Interest' and 'Seal of Approval'. The pattern is loose, and the steering text does not show a clear link.

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

    responsible for image formation by lenses and the eye.⏎- Optical density is the ability of a medium to refract light.⏎When a light ray passes from a substance of one density into a substance of different density, the speed of light changes and bending

  2. #2 peak

    of Oregon and Washington. Let's hope Cascadians welcome them back as the agents of ecological restoration that they are.⏎Posted by Eric de Place | Permalink<|endoftext|>This article is not intended to disparage the

  3. #3 peak

    intended to replace human translation. Neither BioOne nor the owners and publishers of the content make, and they explicitly disclaim, any express or implied representations or warranties of any kind, including, without limitation, representations and warranties as to the functionality of

  4. #4 peak

    up to 40,000 years. This is because the photons have to travel about twice the distance from the Earth to the Moon to<|endoftext|>Posted by Admin on August, 13, 2021⏎Every flame

  5. #5 peak

    . The adverb "yesterday" makes it clear that the action was in the past. The rules of grammar, which require the past tense of the verb plus the word "yesterday, make it redundantly clear that the

  6. #6 peak

    on some historic maps<|endoftext|>[File photo: DW]⏎A statue of Karl Marx designed by a Chinese artist was accepted on March 13 by the city council of Trier, a German town where the revolutionary socialist thinker was born. The statute

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 future of the human mind. What is the future of the human brain in terms of capability? and what is the future of the human brain in terms of capability? and what is the future of the human brain in terms of capability?

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

    5i2.5⏎© 2018 Author. This work is licensed under the Creative Commons Attribution 4.0 International License.⏎Conflicts of Interest: None⏎Mild traumatic brain injury is defined as a traumatically

  2. #2 peak

    that solve the mathematical equations and algorithms which simulate the pollutant dispersion...⏎. To the general public, however, the AMS is best known for its "Seal of Approval" to television and radio meteorologists.⏎The AMS publishes nine atmospheric and related

  3. #3 peak

    - Custodial care: Personal care that's often given by family caregivers, nurses' aides, or home health workers who provide assistance with what are called "activities of daily living" such as bathing, eating, and dressing.⏎Long-term care is

  4. #4 peak

    cesis.⏎For Nietzsche, destiny keeps the form of Amor fati (Love of Fate) through the important element of Nietzsche's philosophy, the "will to power" (der Wille zur Macht), the basis of human behavior,

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. L11:1246 excites
  2. L9:2392 suppresses
  3. L8:608 excites
  4. L11:2016 excites
  5. L9:1026 suppresses
  6. L11:416 weakly suppresses
  7. L10:952 weakly excites
  8. L5:1392 weakly excites
  9. L10:132 weakly excites
  10. L8:1907 weakly suppresses
  11. L8:2722 weakly suppresses
  12. L8:1528 weakly suppresses
  13. L6:89 weakly suppresses
  14. L10:2105 weakly suppresses
  15. L6:460 weakly excites
  16. L0:3059 weakly suppresses
  17. L1:2623 weakly suppresses
  18. L6:2648 weakly excites "the" in nickname titles
  19. L8:1314 weakly excites
  20. L11:1392 weakly excites 'of' in fixed technical phrases
  21. L11:133 weakly suppresses
  22. L9:1633 weakly excites
  23. L5:2965 weakly excites
  24. L7:1684 weakly excites

Feeds later layers

  1. L14:1737 suppresses
  2. L16:2715 suppresses
  3. L14:677 excites
  4. L14:759 suppresses
  5. L15:646 suppresses
  6. L15:227 suppresses
  7. L13:1335 suppresses 'of' inside fixed noun phrases
  8. L17:870 suppresses
  9. L14:1265 suppresses
  10. L15:435 excites
  11. L14:491 excites
  12. L18:192 excites
  13. L16:434 suppresses
  14. L17:375 excites
  15. L13:1929 suppresses
  16. L13:1513 suppresses
  17. L19:2009 suppresses
  18. L13:364 excites
  19. L13:1181 suppresses 'of' after a definition or noun
  20. L15:42 suppresses
  21. L16:498 excites
  22. L13:2078 suppresses
  23. L13:416 suppresses
  24. L14:1082 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 -2.23 and -2.131 values between -2.04 and -1.941 values between -1.76 and -1.661 values between -1.57 and -1.471 values between -1.19 and -1.101 values between -1.10 and -1.001 values between -1.00 and -0.911 values between -0.91 and -0.817 values between -0.72 and -0.623 values between -0.62 and -0.5314 values between -0.53 and -0.4441 values between -0.44 and -0.3496 values between -0.34 and -0.25253 values between -0.25 and -0.15624 values between -0.15 and -0.061405 values between -0.06 and 0.04835 values between 0.04 and 0.13425 values between 0.13 and 0.23130 values between 0.23 and 0.3242 values between 0.32 and 0.4112 values between 0.41 and 0.515 values between 0.51 and 0.602 values between 0.60 and 0.701 values between 0.70 and 0.794 values between 0.79 and 0.891 values between 1.45 and 1.55offnegativepositiveoften

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. anni
  2. cies
  3. ificantly
  4. 八个
  5. 应在
  6. しい
  7. 八十
  8. setParameter
  9. evacuate
  10. isphere

Suppresses

  1. agon
  2. ester
  3. าก
  4. entral
  5. emies
  6. вор
  7. oten
  8. máu
  9. .pol
  10. الذ