Neuron Explorer

Qwen3-0.6B

the 'de' syllable starting 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 fires on the piece 'de' at the start of a word fragment, as in dehydrated, DeGrasse, devious, decongestants and deities. Steering up made the model drift into text about small things and feelings, and pushing down made it repeat the word brand.

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

    India is indeed hot and cruel, the long hot and often humid summer days, and the even more uncomfortable sultry, hot evenings in the capital leave us all dehydrated and exhausted. The heat, and the air conditioning, all take a

  2. #2 peak

    chances in<|endoftext|>About This Episode⏎What were early humans like? What did they eat? What did they do? On this episode of StarTalk Radio, Neil deGrasse Tyson and comic co-host Matt Kirshen investigate the diet of early

  3. #3 peak

    expression “talking about the birds and the bees” represents the most common version of our attempts to do this with some degree of biological accuracy. The jokes and devious explanations used by some parents also acknowledge our discomfort in dealing with the subject.⏎

  4. #4 peak

    body from stress – Our body needs to be de-stressed once in a while. Long work hours, unhealthy diet, and irregular

  5. #5 peak

    each step should be at least 10 inches, but no more than 12⏎Frequently Asked Questions⏎How Do You Build Stairs Between Two Decks?⏎There are a few ways to build stairs between decks, but the most

  6. #6 peak

    time to fix it, a lot of time,” he told MEE.⏎Among the shattered Assyrian treasures are two huge Lamassu (winged bull protective deities) from the ancient

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: brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand

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

    Snow White,” the beloved, first, full-length animated feature film, created by the Walt Disney Studios, is celebrating its 75th anniversary this year with a major exhibit at the Walt Disney Family Museum in San Francisco.⏎Called “Snow White

  2. #2 peak

    was occupied during the time of Moses and the Biblical Exodus.2 What do we make of this?⏎In “Kadesh-Barnea—In the Bible and on the Ground” in the September/October

  3. #3 peak

    algae greenhouses and the processing of algae – a plant with future.<|endoftext|>“I was forty years old when Moses the servant of the Lord sent me from Kadesh-Barnea to spy out the land; and I brought him an honest report.”⏎Where

  4. #4 peak

    desires, the right kind method to pursue happiness, and the correct plan for an excellent lifestyles.⏎Read or Download Aristotle for Everybody: Difficult Thought Made Easy PDF⏎Best greek &

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. L24:637 excites
  2. L21:2130 suppresses
  3. L20:2890 suppresses
  4. L25:665 weakly suppresses
  5. L23:792 weakly excites
  6. L21:2084 weakly suppresses
  7. L17:1892 weakly excites
  8. L19:2971 weakly excites
  9. L19:1272 weakly suppresses
  10. L21:2743 weakly suppresses
  11. L20:180 weakly suppresses
  12. L10:1610 weakly excites
  13. L24:2875 weakly excites
  14. L24:2370 weakly excites
  15. L0:674 weakly suppresses
  16. L5:457 weakly suppresses 'a' after 'tell', 'bring' or 'much of'
  17. L16:126 weakly excites
  18. L4:137 weakly excites
  19. L20:901 weakly suppresses
  20. L21:1790 weakly excites
  21. L21:7 weakly excites mixed, weak code and reference hits
  22. L18:2744 weakly excites
  23. L16:139 weakly suppresses
  24. L1:1833 weakly excites

Feeds later layers

  1. L27:1658 weakly excites
  2. L27:1027 weakly suppresses
  3. L27:2611 weakly excites
  4. L27:2361 weakly suppresses
  5. L27:880 weakly excites
  6. L27:1628 weakly suppresses
  7. L27:2597 weakly suppresses
  8. L27:744 weakly excites
  9. L27:2201 weakly suppresses
  10. L27:2838 weakly suppresses
  11. L27:625 weakly excites
  12. L27:2848 weakly excites
  13. L27:612 weakly excites
  14. L27:738 weakly excites
  15. L27:925 weakly suppresses
  16. L27:2080 weakly excites first piece of a split name
  17. L27:1158 weakly excites
  18. L27:1587 weakly suppresses
  19. L27:681 weakly excites
  20. L27:850 weakly excites
  21. L27:185 weakly suppresses
  22. L27:1567 weakly suppresses
  23. L27:2072 weakly suppresses
  24. L27:523 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 -27.03 and -25.631 values between -21.42 and -20.021 values between -20.02 and -18.611 values between -13.00 and -11.603 values between -11.60 and -10.194 values between -10.19 and -8.796 values between -8.79 and -7.399 values between -7.39 and -5.9815 values between -5.98 and -4.5831 values between -4.58 and -3.1878 values between -3.18 and -1.78766 values between -1.78 and -0.372618 values between -0.37 and 1.03213 values between 1.03 and 2.4379 values between 2.43 and 3.8434 values between 3.84 and 5.2417 values between 5.24 and 6.6410 values between 6.64 and 8.0510 values between 8.05 and 9.453 values between 9.45 and 10.853 values between 10.85 and 12.262 values between 12.26 and 13.661 values between 15.06 and 16.471 values between 16.47 and 17.871 values between 27.69 and 29.09offnegativepositiveoften

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. [++
  2. .annotations
  3. 篾
  4. glue
  5. 注
  6. edis
  7. edes
  8. s
  9. 副局长
  10. 独角

Suppresses

  1. brand
  2. Brand
  3. Brand
  4. BRAND
  5. .brand
  6. -brand
  7. brand
  8. 品牌的
  9. _brand
  10. brands