Neuron Explorer

Qwen3-0.6B

'something' or 'anything' after a verb

A neuron in layer 21 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 'something' and 'anything' that come after a verb, as in 'learn something' or 'discovered anything'. When it was turned up, the model kept repeating the word 'important'.

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

    to the phone or to other people. Now, people never have to be alone and even when they are with other people, there is still this fear that maybe something

  2. #2 peak

    - they don't twinkle, they shine," Johnstone says. He spends as much time as possible gazing at the skies: "I haven't discovered anything of importance yet, like maybe a comet or dwarf planet, but I'm still

  3. #3 peak

    it doesn’t mean that the students from other classes can’t read this article. Of course, they can.⏎You’ll surely learn something from these moral stories at the end.⏎Read some other English moral stories.⏎H

  4. #4 peak

    . A solar storm would still pose a risk for wandering astronauts, he adds, but there would be time to return to a shielded base.⏎To continue reading this article, subscribe to receive access to all of newscientist.com, including

  5. #5 peak

    yming with aight in the 1980s. Why doesn’t the OED cite any of them? Have the Oxford boffins got something against DJ Jazzy Jeff & The Fresh Prince? Maybe, but I suspect

  6. #6 peak

    is crucial for creativity.⏎Do not be afraid to hear others as you compose. There are plenty of different views on writing, and you’re probably going to learn something as you write. Simply listening to what others say may allow you<|endoftext|>It is

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 5th and 6th of 7th 8th 9th 10th 11th 12th 12th 12nd 12nd 12nd 1

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

    of on-line web content, which is often made available to customers using the web. This write-up briefly covers the different types of posting as well as goes over just how these different approaches differ from each various other and just how they can be useful

  2. #2 peak

    Terms Tests are given every six weeks.⏎Many of the literature selections are taken from Norton Anthologies although they can also be found online. Additionally, one or more whole book studies are covered in each course. For the British Literature course, the author

  3. #3 peak

    ek's father, Sabu, did not approve of their relationship as he knew that Aya's dream was to settle in Alexandria, and he feared that her mixed heritage would corrupt his teachings.⏎Later

  4. #4 peak

    hind legs, facing towards the light – reaffirming that Pine Martens truly are superstar performers in front of camera traps.⏎The spread of Pine Martens across much of the Scottish mainland north of Edinburgh and Glasgow is a success story for species resilience

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. L19:2926 suppresses
  2. L20:1012 excites
  3. L19:2219 excites
  4. L19:3017 suppresses
  5. L19:2122 suppresses
  6. L18:801 excites
  7. L20:809 excites
  8. L18:1596 suppresses
  9. L18:2398 excites
  10. L18:482 suppresses
  11. L5:2687 weakly suppresses
  12. L14:2609 weakly excites
  13. L19:2843 weakly suppresses
  14. L17:2159 weakly suppresses
  15. L14:1519 weakly excites
  16. L19:2149 weakly excites
  17. L5:1410 weakly excites
  18. L6:1755 weakly suppresses
  19. L18:2058 weakly suppresses
  20. L19:2274 weakly excites
  21. L17:1086 weakly excites 'something' as an object
  22. L19:914 weakly excites
  23. L13:2736 weakly excites
  24. L9:1663 weakly suppresses

Feeds later layers

  1. L27:2019 excites
  2. L24:568 excites
  3. L22:1297 excites
  4. L24:915 excites
  5. L23:530 excites
  6. L23:1323 excites
  7. L25:1476 excites
  8. L22:2814 excites
  9. L23:283 excites
  10. L25:1891 excites
  11. L23:1726 excites
  12. L22:415 suppresses
  13. L26:1849 excites
  14. L22:901 weakly excites
  15. L25:339 weakly excites
  16. L23:511 weakly excites
  17. L27:460 weakly excites
  18. L23:143 weakly suppresses
  19. L24:372 weakly excites
  20. L22:691 weakly excites
  21. L22:2326 weakly excites the word 'well' before -known
  22. L27:620 weakly excites
  23. L26:1612 weakly excites
  24. L23:2291 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.

3 values between -6.57 and -6.201 values between -5.82 and -5.454 values between -5.45 and -5.073 values between -5.07 and -4.699 values between -4.69 and -4.321 values between -4.32 and -3.946 values between -3.94 and -3.575 values between -3.57 and -3.198 values between -3.19 and -2.8114 values between -2.81 and -2.4426 values between -2.44 and -2.0626 values between -2.06 and -1.6944 values between -1.69 and -1.3181 values between -1.31 and -0.93133 values between -0.93 and -0.56657 values between -0.56 and -0.182004 values between -0.18 and 0.20520 values between 0.20 and 0.57140 values between 0.57 and 0.9561 values between 0.95 and 1.3234 values between 1.32 and 1.7021 values between 1.70 and 2.0824 values between 2.08 and 2.4519 values between 2.45 and 2.8314 values between 2.83 and 3.208 values between 3.20 and 3.589 values between 3.58 and 3.967 values between 3.96 and 4.336 values between 4.33 and 4.715 values between 4.71 and 5.082 values between 5.08 and 5.463 values between 5.46 and 5.842 values between 5.84 and 6.211 values between 6.21 and 6.591 values between 6.59 and 6.962 values between 6.96 and 7.342 values between 7.34 and 7.721 values between 7.72 and 8.091 values between 8.09 and 8.47offnegativepositiveoften

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. important
  2. 重要
  3. 重要的
  4. important
  5. importante
  6. importantes
  7. wicht
  8. важ
  9. valuable
  10. 的重要

Suppresses

  1. imated
  2. odore
  3. 其它问题
  4. ลบ
  5. olson
  6. 投资基金
  7. arella
  8. 仅代表
  9. Erect
  10. 彘