Neuron Explorer

Qwen3-0.6B

artificial and robotic in tech text

A neuron in layer 20 of Qwen3-0.6B. Its strongest snippets are fairly consistent: several of them peak on the same token. It is featured, so it has steering outputs.

What this neuron does

This neuron fires on words like 'artificial' and 'robotic' when the text is about AI, robots or new technology. Its negative side fires on 'on' in phrases about lying or walking on a surface, like 'lying on their back'. Turning it up gave replies that didn't stay on the AI 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

    could not interact with humans.⏎Artificial Intelligence is Different⏎In recent years, big data has revolutionized the world.

  2. #2 peak

    laws on their own. On the higher end, they could operate autonomously to conduct arrests of suspects who might resist arrest violently. Near the highest end would be robotic police at least as capable as human beings.⏎While there are legitimate worries that police

  3. #3 peak

    behavior," and means that the researchers are able to reduce error rates by nearly a third.⏎Here's more on that concept from a June Microsoft Research post:⏎Artificial neural networks are mathematical models of low-level circuits in the human brain. They have

  4. #4 peak

    for science and cryptography.⏎In the 21st century, we find ourselves on the cusp of a technological revolution unlike anything humanity has ever witnessed. From artificial intelligence (AI) to blockchain, quantum computing to biotechnology, emerging technologies are

  5. #5 peak

    blockchain, quantum computing, and biotechnology. It discusses their impact on various sectors and emphasizes the need for responsible adoption and ethical considerations.⏎- Artificial intelligence’s ability to learn and adapt is revolutionizing

  6. #6 peak

    -friendly GUIs akin to those seen on early Macs.⏎Near the end of his talk, Jobs looked even further into the future, hinting that technology like artificial intelligence and predictive computing could one day be a possibility.⏎He recalls how books helped

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 have a hand on my back and I am on a bed on my back on the bed. I am on my stomach on a bed on my back. I am on my hands and feet on a bed on my back. I have

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

    the zoo's clinic.⏎Workers at the zoo say that's when she started walking upright exclusively, reports CBC News.⏎Monkeys usually alternate between upright walking and moving on all fours.⏎A zoo veterinarian says he's not sure why she has altered

  2. #2 peak

    him/her and will learn to crawl, stand and eventually walk.⏎Why is tummy time important?⏎Infants spend a lot of time lying on their back, especially a few weeks after being born. If they’re not lying

  3. #3 peak

    in foals include intermittent colic or diarrhea, rough hair coat, poor growth, or drooling. Discomfort may be shown by grinding the teeth, lying on the back to try to ease abdominal pain, or interrupted nursing. The foal

  4. #4 peak

    should make sure that your infant is at least getting 15-20 minutes of tummy time during the day. Once your child is comfortable with lying on their tummy, you can help them build their strength and motor skills via different

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. L16:2955 suppresses
  2. L11:1148 excites
  3. L8:1948 excites
  4. L11:626 suppresses
  5. L7:1851 suppresses
  6. L4:206 suppresses
  7. L9:722 weakly excites
  8. L15:1986 weakly excites
  9. L4:2687 weakly excites
  10. L9:2151 weakly suppresses
  11. L9:2526 weakly suppresses
  12. L9:2564 weakly excites
  13. L18:1015 weakly excites
  14. L14:1947 weakly excites
  15. L9:1374 weakly excites
  16. L6:2059 weakly suppresses
  17. L10:2248 weakly suppresses
  18. L16:1936 weakly suppresses
  19. L15:2694 weakly excites
  20. L10:1442 weakly excites
  21. L11:1530 weakly suppresses
  22. L9:517 weakly suppresses
  23. L18:208 weakly suppresses
  24. L2:2838 weakly excites 'Head Start' and 'follow up'

Feeds later layers

  1. L23:1677 weakly excites
  2. L27:246 weakly excites
  3. L27:381 weakly suppresses
  4. L22:2295 weakly excites
  5. L22:2318 weakly excites
  6. L22:499 weakly suppresses
  7. L24:332 weakly suppresses
  8. L26:929 weakly excites
  9. L22:38 weakly excites
  10. L24:1796 weakly suppresses
  11. L23:1657 weakly excites
  12. L23:1849 weakly excites
  13. L23:2899 weakly excites
  14. L23:2037 weakly excites
  15. L22:2588 weakly suppresses
  16. L22:1730 weakly excites
  17. L21:1868 weakly suppresses
  18. L25:3023 weakly suppresses medical specialists and eye doctors
  19. L26:951 weakly suppresses
  20. L21:2257 weakly suppresses
  21. L21:2203 weakly excites
  22. L24:2465 weakly suppresses
  23. L24:1560 weakly suppresses
  24. L26:845 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.

2 values between -14.48 and -13.841 values between -13.84 and -13.211 values between -13.21 and -12.571 values between -12.57 and -11.941 values between -10.03 and -9.392 values between -9.39 and -8.761 values between -8.76 and -8.122 values between -8.12 and -7.491 values between -7.49 and -6.857 values between -6.85 and -6.222 values between -6.22 and -5.584 values between -5.58 and -4.953 values between -4.95 and -4.3110 values between -4.31 and -3.6818 values between -3.68 and -3.0421 values between -3.04 and -2.4135 values between -2.41 and -1.7786 values between -1.77 and -1.13203 values between -1.13 and -0.502247 values between -0.50 and 0.141109 values between 0.14 and 0.7773 values between 0.77 and 1.4135 values between 1.41 and 2.0417 values between 2.04 and 2.6810 values between 2.68 and 3.315 values between 3.31 and 3.953 values between 3.95 and 4.581 values between 4.58 and 5.223 values between 5.22 and 5.861 values between 6.49 and 7.131 values between 7.13 and 7.761 values between 9.67 and 10.301 values between 10.30 and 10.94offnegativepositiveoften

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. coli
  2. soát
  3. nage
  4. ichier
  5. music
  6. NCY
  7. 妮
  8. hatt
  9. irting
  10. NDAR

Suppresses

  1. 枕
  2. bed
  3. ammed
  4. ']}'
  5. odiac
  6. บน
  7. 醒来
  8. foss
  9. offs
  10. 床上