Neuron Explorer

Qwen3-0.6B

repeated words and ellipsis dots

A neuron in layer 14 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

It reacts to a word or punctuation mark that repeats or stutters, like 'very, very' or the dots in '. . .'. Turning it up did not produce a clear matching theme in the model's writing, and turning it down caused repetition.

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

    agreed to assume claims by citizens of the United States against Spain. In Article III, the treaty stated that: "The boundary line between the two countries. . . shall begin on the Gulph [sic] of Mexico, at the mouth of

  2. #2 peak

    , 2003, the National⏎Transportation Safety Board (NTSB) made its boldest statement to date.⏎They stated that “ . . . the safe operation of fifteen-passenger⏎vans requires a knowledge and skill level

  3. #3 peak

    some 44,000 generations, equivalent to some million years of generations of a human population (but the evolutionary opportunities for humans would be far, far less, due to the small population numbers limiting the number of mutational possibilities;

  4. #4 peak

    is defined as to:⏎“engage in a course of vexatious conduct or to make a demand or a sexual solicitation or advance that one knows or ought reasonably to know is unwelcome.”⏎Severe Act or Pattern of Actions⏎Let’s

  5. #5 peak

    ! How changed she was! How very, very different she looked from the last time she had seen herself in a big mirror!

  6. #6 peak

    in another(),so this prints 20. Thus final answer is 10 0 20. I hope i have been clear enough. ------------------ Come on in !! Drinks are on the house in the Big Moose Saloon !

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 been thinking about the need to be more engaged, and I have been thinking about the need to be more engaged, but I have been thinking about the, I have been, a, I, I, an, I, I

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 business.⏎This week’s article by Jeremy Rosen on Rav Lichtenstein is available from http://jeremyrosen.blogspot.co.uk/2015/08/rav-lichtenstein-zl.html⏎Your comment will be posted

  2. #2 peak

    the issue of Southern autonomy.⏎In 1924, the 28th president of the United States, Woodrow Wilson, died in Washington, D.C., at age 67.⏎In 1930, the

  3. #3 peak

    NR in decibels. From Step 1, we have the decibel value LdB as LdB = 10 log10 (P1/P0). Now let P1 = MAX^2 and P0 = MSE.

  4. #4 peak

    Pesticides in Produce⏎Researchers Highlight Pesticides in Produce,⏎Baby Food, Tap Water⏎Washington, D.C. – Environmental Working Group has released the eighth edition of its Shopper’s

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. L9:7 excites
  2. L13:1006 excites
  3. L10:33 suppresses
  4. L13:755 suppresses
  5. L13:351 weakly suppresses
  6. L11:173 weakly suppresses
  7. L6:4 weakly suppresses
  8. L10:527 weakly excites
  9. L13:1401 weakly suppresses
  10. L13:1428 weakly excites
  11. L13:494 weakly excites
  12. L6:0 weakly excites
  13. L13:2687 weakly excites
  14. L10:3035 weakly suppresses
  15. L0:700 weakly suppresses
  16. L10:394 weakly excites
  17. L3:3 weakly suppresses
  18. L9:746 weakly suppresses
  19. L10:1216 weakly suppresses
  20. L1:1 weakly excites
  21. L5:394 weakly suppresses
  22. L5:872 weakly suppresses 'or' introducing an alternate name
  23. L11:966 weakly excites
  24. L13:287 weakly excites

Feeds later layers

  1. L17:1590 suppresses
  2. L22:269 weakly suppresses
  3. L20:3 weakly excites
  4. L18:501 weakly suppresses
  5. L18:574 weakly excites
  6. L15:78 weakly excites
  7. L19:1747 weakly suppresses
  8. L19:1058 weakly excites
  9. L25:1668 weakly suppresses
  10. L17:560 weakly suppresses
  11. L26:274 weakly excites
  12. L16:1943 weakly excites
  13. L17:2893 weakly suppresses
  14. L18:2055 weakly excites
  15. L17:2977 weakly suppresses
  16. L17:408 weakly suppresses
  17. L21:2623 weakly suppresses
  18. L27:1748 weakly suppresses
  19. L15:122 weakly excites
  20. L21:2616 weakly suppresses
  21. L16:1639 weakly suppresses
  22. L20:1526 weakly suppresses
  23. L18:546 weakly suppresses
  24. L17:1622 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 -1.63 and -1.542 values between -1.35 and -1.261 values between -1.26 and -1.163 values between -1.16 and -1.073 values between -1.07 and -0.975 values between -0.97 and -0.885 values between -0.88 and -0.789 values between -0.78 and -0.6915 values between -0.69 and -0.6023 values between -0.60 and -0.5024 values between -0.50 and -0.4168 values between -0.41 and -0.31154 values between -0.31 and -0.22323 values between -0.22 and -0.12822 values between -0.12 and -0.031543 values between -0.03 and 0.07426 values between 0.07 and 0.16206 values between 0.16 and 0.25101 values between 0.25 and 0.3559 values between 0.35 and 0.4432 values between 0.44 and 0.5426 values between 0.54 and 0.6318 values between 0.63 and 0.7313 values between 0.73 and 0.827 values between 0.82 and 0.914 values between 0.91 and 1.013 values between 1.01 and 1.102 values between 1.10 and 1.201 values between 1.20 and 1.292 values between 1.29 and 1.392 values between 1.39 and 1.481 values between 1.67 and 1.761 values between 1.86 and 1.951 values between 1.95 and 2.052 values between 2.05 and 2.14offnegativepositiveoften

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. bler
  2. ям
  3. 两者
  4. 者
  5. .TryParse
  6. (kind
  7. 研究所
  8. 的是
  9. 两种
  10. лась

Suppresses

  1. ainers
  2. -ahead
  3. GridColumn
  4. akening
  5. Ahead
  6. .nextSibling
  7. 玒
  8. AppBundle
  9. 鲠
  10. Posts