Neuron Explorer

Qwen3-0.6B

"over" in "all over the world"

A neuron in layer 1 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 "over" when it comes right after "all", as in "all over the world". Turning it up made the model's text drift into unrelated topics like code and random puzzles, so the steering did not show a clear theme.

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

    and distributed to classrooms all over the world. They also offer workshops to assist K-12 teachers interested in

  2. #2 peak

    jet stream and producing more ferocious hurricanes and floods all over the world. But its effects will be acutely felt here, according to a

  3. #3 peak

    s" in some of the oddest of places such as the tops of power line poles. They have been seen in Chicago, New York as well as all over South Florida. Speculation as to how these feral flocks developed seems topping

  4. #4 peak

    video is not just about having fun, it brings people together from all over the world.⏎With the right music, it now becomes an inspirational video that can

  5. #5 peak

    , engineers will be able to construct microthin digital displays that can be<|endoftext|>Diabetes mellitus is a non communicable disease but is spreading like a pandemic all over the world. Each and every house is having at least two members who are diagnosed

  6. #6 peak

    Patch, consists of waste such as toilet seats, camera cases, and oil drums, floating in the middle of the Pacific ocean. The trash accumulates from all over the world and as it moves into the ocean, it becomes trapped at the center

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: "In a 2 game, the two players are A and B. If A has a strategy of 1, what should the other player do?" (A) and (B) to respond? The other player has a strategy of

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

    R. L., Sobel, R. S., & Macpherson, D. A. (2018). Economics: Private⏎and public choice (16

  2. #2 peak

    be found professionals and graduates of British and American universities.⏎A few trade unions had existed for the protection of workers of certain categories (e.g., the Nigeria Civil Service Union) before trade unionism became legalised in 1938.

  3. #3 peak

    makes it easier for<|endoftext|>The latest weapon unleashed to battle California's growing energy demand comes in the form of free software. The State of California Energy Commission's Public Interest

  4. #4 peak

    also read the full article published in Transactions of the Institute of British Geographers here.⏎Ethnic diversity in English city neighbourhoods has been the subject of longstanding public interest. While some have argued that diversity is associated with lower social cohesion and trust,

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. L0:1301 suppresses
  2. L0:1224 suppresses
  3. L0:178 suppresses
  4. L0:1 weakly suppresses
  5. L0:1229 weakly excites
  6. L0:512 weakly suppresses
  7. L0:179 weakly suppresses
  8. L0:571 weakly excites
  9. L0:1638 weakly excites
  10. L0:2351 weakly excites
  11. L0:732 weakly excites
  12. L0:2436 weakly suppresses
  13. L0:1660 weakly excites
  14. L0:2408 weakly excites
  15. L0:2645 weakly excites
  16. L0:1167 weakly suppresses
  17. L0:2089 weakly suppresses
  18. L0:943 weakly excites
  19. L0:12 weakly suppresses
  20. L0:187 weakly excites
  21. L0:1554 weakly excites
  22. L0:2522 weakly excites
  23. L0:1075 weakly excites
  24. L0:209 weakly excites

Feeds later layers

  1. L13:2627 weakly suppresses
  2. L8:2160 weakly excites
  3. L21:1547 weakly suppresses
  4. L9:2479 weakly suppresses
  5. L3:411 weakly suppresses
  6. L12:2502 weakly suppresses
  7. L4:2276 weakly suppresses
  8. L8:3065 weakly suppresses
  9. L2:2399 weakly suppresses
  10. L25:400 weakly suppresses
  11. L22:1673 weakly suppresses
  12. L7:1926 weakly suppresses
  13. L27:2658 weakly suppresses
  14. L21:2097 weakly excites
  15. L26:2117 weakly suppresses
  16. L13:3067 weakly suppresses
  17. L2:1868 weakly excites
  18. L2:2591 weakly suppresses
  19. L3:2420 weakly excites
  20. L11:2267 weakly suppresses
  21. L2:2726 weakly excites 'by' after 'characterized' or 'marked'
  22. L4:2456 weakly suppresses
  23. L9:2810 weakly excites
  24. L5:1691 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 -0.89 and -0.811 values between -0.57 and -0.492 values between -0.49 and -0.414 values between -0.41 and -0.3327 values between -0.33 and -0.2552 values between -0.25 and -0.17218 values between -0.17 and -0.091057 values between -0.09 and -0.011954 values between -0.01 and 0.07435 values between 0.07 and 0.15111 values between 0.15 and 0.2331 values between 0.23 and 0.306 values between 0.30 and 0.385 values between 0.38 and 0.461 values between 0.46 and 0.541 values between 0.62 and 0.701 values between 1.49 and 1.571 values between 2.21 and 2.29offnegativepositiveoften

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. erable
  2. ически
  3. ome
  4. omes
  5. lobal
  6. Respect
  7. ическое
  8. ейств
  9. ened
  10. ergy

Suppresses

  1. bull
  2. -zA
  3. /pub
  4. >*</
  5. |_|
  6. 专区
  7. >'+⏎
  8. _('
  9. mutation
  10. $/,⏎