Neuron Explorer

Qwen3-0.6B

'come' in phrases like 'come across'

A neuron in layer 23 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 lights up on the word 'come' when it is part of a phrase, such as 'come across', 'come up with' or 'come to'. It is switched off by 'came' and 'person' in some cases. Turning it up gave replies that were loosely about thinking or coming to a conclusion, but the link is only partly clear.

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

    hand to genitals), and more! Because there isn’t just one type of sex, there are different options for safer sex. These are some terms that you might come across when we talk about safer sex.⏎Antiretroviral

  2. #2 peak

    out how you came to be on this earth and who lived before you in the family tree.⏎The most common type of DNA testing for genealogy that you’ll come across is an autosomal DNA test. Not only do they tell you your DNA

  3. #3 peak

    be difficult to come up with a great melody starting only with lyrics.⏎Maintain a collection of lyr

  4. #4 peak

    we don’t keep the end in mind. For example, a teen working to save $50 to have spending money for entertainment will more easily be able to come up with a plan than if they were trying to save up $5,0

  5. #5 peak

    there was no consensus on the definition of a microservice and what it actually means. It is quite a buzzword these days but none of the panelists could come to a common understanding. This raised a few questions in my head. Disappointed

  6. #6 peak

    asking them to download the free Aurasma app and bring their device. We talk about the importance of "depth" in education. I'm sure you'll come up with lots of ideas of your own. A. Media:

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: walking walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk walk

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

    ’s This?⏎The game went to penalty kicks and, right when Messi was about to take one, Uncle Simon, dressed in his usual beehive wig, came in with a tray of canapés and asked, ‘Who’s playing?’

  2. #2 peak

    problem solver and go out of your way to help co-workers. You do your job and make it all look easy. Despite your dedication, one day a new person is hired. They have the same education, experience, and skills that you have

  3. #3 peak

    in the Chaos⏎The perils of plastics in the ocean are threatening to the ecosystem, but all is not gloomy. This is where BLUE BEAUTY comes to the rescue. It’s a sub-section of clean or green beauty but domin

  4. #4 peak

    these machines that can capture an underlying spirit, or an underlying set of principles, or an underlying way of looking at the world so that then when the next Aristotle comes around?" Jobs said. "Maybe if he carries around one these machines with him

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. L21:2728 suppresses
  2. L12:1826 weakly suppresses
  3. L20:2514 weakly suppresses
  4. L15:1019 weakly suppresses
  5. L17:217 weakly suppresses
  6. L12:1926 weakly excites
  7. L16:1853 weakly suppresses
  8. L6:1205 weakly suppresses
  9. L20:380 weakly excites
  10. L21:1221 weakly suppresses
  11. L0:2541 weakly suppresses
  12. L7:1851 weakly suppresses
  13. L5:2324 weakly excites
  14. L18:1407 weakly excites
  15. L22:151 weakly suppresses
  16. L4:360 weakly suppresses
  17. L7:2166 weakly excites
  18. L6:465 weakly suppresses
  19. L22:2909 weakly excites
  20. L17:1469 weakly excites
  21. L15:2109 weakly excites
  22. L3:90 weakly suppresses
  23. L0:450 weakly suppresses
  24. L15:1751 weakly suppresses

Feeds later layers

  1. L24:1219 suppresses
  2. L26:1965 suppresses
  3. L25:1027 weakly suppresses
  4. L25:194 weakly excites
  5. L24:1622 weakly excites
  6. L26:3039 weakly suppresses
  7. L27:824 weakly suppresses
  8. L25:2115 weakly suppresses
  9. L27:1008 weakly excites
  10. L24:691 weakly suppresses
  11. L24:411 weakly excites
  12. L27:1050 weakly excites
  13. L24:1301 weakly suppresses
  14. L27:2634 weakly excites
  15. L26:1104 weakly excites
  16. L25:2236 weakly suppresses
  17. L27:2474 weakly suppresses
  18. L27:2129 weakly excites
  19. L27:235 weakly excites
  20. L25:1943 weakly suppresses
  21. L27:1381 weakly excites
  22. L27:1526 weakly suppresses
  23. L24:876 weakly suppresses
  24. L25:754 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 -22.31 and -21.341 values between -16.50 and -15.531 values between -12.62 and -11.651 values between -11.65 and -10.683 values between -9.71 and -8.751 values between -5.84 and -4.874 values between -4.87 and -3.904 values between -3.90 and -2.9310 values between -2.93 and -1.9647 values between -1.96 and -0.992292 values between -0.99 and -0.021461 values between -0.02 and 0.9542 values between 0.95 and 1.9217 values between 1.92 and 2.893 values between 2.89 and 3.854 values between 3.85 and 4.822 values between 4.82 and 5.791 values between 5.79 and 6.762 values between 6.76 and 7.731 values between 7.73 and 8.702 values between 8.70 and 9.671 values between 10.64 and 11.612 values between 12.58 and 13.552 values between 13.55 and 14.522 values between 14.52 and 15.481 values between 15.48 and 16.45offnegativepositiveoften

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. up
  2. _Config
  3. Away
  4. Away
  5. up
  6. لين
  7. away
  8. -up
  9. Guth
  10. away

Suppresses

  1. walk
  2. walk
  3. 走了
  4. walks
  5. walked
  6. 走
  7. walking
  8. entar
  9. ⾛
  10. 步