Neuron Explorer

Qwen3-0.6B

'the' after 'member of' or 'one of'

A neuron in layer 7 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 'the' when it comes after phrases like 'one of' or 'members of', where a group is about to be named. Turning it up gave repetitive or odd text with no clear link to the group idea.

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

    John Polidori. To pass the time, the group read a volume of ghost stories aloud, at which point Byron posed a challenge in which each member of the group would attempt to write such a tale.⏎“The chronology that’s in most

  2. #2 peak

    Baltimore and Washington, D.C., as well as in Florida and some Southern states.⏎Don’t Hebrew Israelites have a community in southern Israel?⏎One part of the wider Black Hebrew Israelite movement began arriving in Israel in the late 19

  3. #3 peak

    attributes. Each node in the tree represents a decision based on a feature, and the tree continues to split the

  4. #4 peak

    be fading away. english⏎and its speakers, owing to their advantage, are going to dominate societies. If members of the elite class in countries where english⏎is not the spoken language, live their lives

  5. #5 peak

    of the very best Old English/Anglo-Saxon poems is followed by footnotes and the translator's comments. Included in the⏎notes are a summary and a detailed analysis of the⏎poem's plot

  6. #6 peak

    house thousands of Japanese American internees. One of the first families to arrive was the Wakatsukis, who were ordered to leave

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 number of, and the number of, and the number of , ... for ... x millionth, which is how much it is for 10% of 24. So the answer is 10% of 2

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

    . In this article, we will look at five strategies for motivating students and keeping their brains engaged over the summer.⏎Engage in Reading Adventures⏎There is no better time to read than during the dog days of summer. Whether it’s sci-fi

  2. #2 peak

    3-million people (more than one-third of the entire U.S. population) to the joyous melodies and excitement of the Fab Four.⏎Nothing would ever be the same

  3. #3 peak

    had been handed down over the years by family and friends and both finally landed in the lap of Blight, who is director of Yale University's Gilder Lehrman Center for the Study of Slavery, Resistance, and Abolition.⏎Even

  4. #4 peak

    update on: 2006-SEP-23 Matthew 12:30, Luke 11:23, and Romans 16:17. Unfortunately, it is the other verses that are more frequently

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. L6:44 suppresses
  2. L6:3 excites
  3. L6:187 suppresses blank token before a number or after a count
  4. L6:93 suppresses
  5. L5:30 suppresses
  6. L5:248 weakly suppresses
  7. L6:376 weakly suppresses
  8. L6:301 weakly suppresses
  9. L1:2157 weakly suppresses
  10. L6:226 weakly excites
  11. L3:2766 weakly suppresses
  12. L6:489 weakly excites
  13. L5:10 weakly excites
  14. L5:2156 weakly suppresses
  15. L6:2146 weakly excites
  16. L1:2866 weakly excites
  17. L2:224 weakly excites
  18. L2:411 weakly excites
  19. L6:10 weakly excites
  20. L1:194 weakly excites
  21. L2:1908 weakly excites
  22. L6:346 weakly suppresses
  23. L4:15 weakly excites
  24. L1:2700 weakly excites

Feeds later layers

  1. L8:1087 weakly excites
  2. L10:17 weakly excites
  3. L8:772 weakly suppresses "it" as a subject pronoun
  4. L17:3007 weakly excites
  5. L11:315 weakly suppresses
  6. L10:97 weakly excites
  7. L9:191 weakly suppresses
  8. L16:2655 weakly suppresses
  9. L10:1448 weakly excites
  10. L11:23 weakly excites
  11. L23:1234 weakly excites word pieces inside names and terms
  12. L11:378 weakly excites
  13. L20:379 weakly suppresses digits inside dates
  14. L12:2186 weakly suppresses
  15. L16:1006 weakly suppresses
  16. L22:1073 weakly suppresses
  17. L24:2308 weakly suppresses
  18. L16:527 weakly excites
  19. L19:1174 weakly suppresses
  20. L21:675 weakly suppresses
  21. L15:779 weakly suppresses
  22. L17:30 weakly excites
  23. L23:2964 weakly suppresses
  24. L8:315 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 -2.21 and -2.023 values between -1.47 and -1.299 values between -1.29 and -1.1119 values between -1.11 and -0.9342 values between -0.93 and -0.7496 values between -0.74 and -0.56212 values between -0.56 and -0.38508 values between -0.38 and -0.201153 values between -0.20 and -0.011188 values between -0.01 and 0.17321 values between 0.17 and 0.35114 values between 0.35 and 0.5440 values between 0.54 and 0.7235 values between 0.72 and 0.9029 values between 0.90 and 1.0930 values between 1.09 and 1.2714 values between 1.27 and 1.4519 values between 1.45 and 1.6313 values between 1.63 and 1.8211 values between 1.82 and 2.009 values between 2.00 and 2.185 values between 2.18 and 2.364 values between 2.36 and 2.556 values between 2.55 and 2.738 values between 2.73 and 2.912 values between 2.91 and 3.107 values between 3.10 and 3.282 values between 3.28 and 3.461 values between 3.46 and 3.643 values between 3.64 and 3.831 values between 3.83 and 4.012 values between 4.01 and 4.191 values between 4.92 and 5.11offnegativepositiveoften

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. @brief
  2. ArrayOf
  3. Handling
  4. InOut
  5. 著
  6. bane
  7. -bl
  8. esch
  9. -li
  10. eros

Suppresses

  1. груз
  2. utta
  3. ата
  4. iden
  5. ��
  6. utory
  7. руз
  8. 意见反馈
  9. ottage
  10. thị