Neuron Explorer

Qwen3-0.6B

'a' meaning one single

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 small word 'a' when it means 'one single thing', as in 'a single chip' or 'a common ancestor'. Turning it up made the model repeat 'single' and 'a single person', so it seems tied to the idea of one.

Written by Claude from this neuron's activations and steering outputs. Confidence: high. 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

    other⏎storage engines. Also, this warning against sharing a data⏎directory among servers always applies in an NFS environment.⏎Permitting multiple MySQL servers to access a common data⏎directory over NFS is a very bad idea. The⏎primary problem

  2. #2 peak

    , what causes depression, it is important to understand that there is a wide range of risk factors that can influence the chances of a person developing depression, rather than a singular cause.⏎Your genetics can be a significant factor in determining how likely you are

  3. #3 peak

    CPU and graphics processor or GPU is generally present separately. Although the possibility technology PC processor to have an integrated graphics processor.⏎The CPU and GPU are generally integrated into a circuit board (motherboard) in which there is a chipset. This Chipset

  4. #4 peak

    that is designed to control specific hardware functions. It consists of a central processing unit (CPU), memory, input/output peripherals, and various interfaces, all integrated into a single chip. Microcontrollers are programmable, which means that developers can write software

  5. #5 peak

    (which form the sums of the product terms). We can simplify this representation still further ( Figure 5.11 c), by drawing the input lines to a multiple-input AND gate as if they were one horizontal wire, which we call a

  6. #6 peak

    .⏎What is a "Microprocessor"?⏎A microprocessor is essentially an entire basic computer fitted on a single chip . Sure, computers purchased today usually come with peripherals like monitors, hard

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: one one one one one one one one one one one one one one one one one one one one one one one one one one one one one one one one one one one one one one one one one one one one one one one one

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

    can safely remove.⏎But remember that you should never cut off more than two-thirds of a root at any time.⏎Also,<|endoftext|>Title VII makes sexual discrimination, i.e., using gender

  2. #2 peak

    obstacles even if we don’t know it, providing our daily portion from His granaries in the mysterious world of the beyond. We are not given too much all at once

  3. #3 peak

    materials such as MyPlate. A "portion" is the amount of a food you choose to eat at any one time — which may be more or less than a serving.⏎Here are some

  4. #4 peak

    if you try to do it all at once, you get overwhelmed," she said.⏎Don't get discouraged if you find

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. L22:1176 strongly excites
  2. L20:1279 strongly suppresses
  3. L17:2592 excites
  4. L18:833 excites
  5. L20:1148 excites
  6. L22:1033 excites
  7. L17:2379 suppresses
  8. L19:2270 suppresses
  9. L18:2927 suppresses
  10. L16:2671 suppresses
  11. L15:1139 suppresses
  12. L19:2358 suppresses "any of" before "these" or "the"
  13. L17:2128 excites
  14. L17:2617 suppresses
  15. L21:1717 excites
  16. L15:2275 suppresses
  17. L19:2468 suppresses
  18. L18:2304 excites
  19. L20:2787 excites
  20. L21:828 suppresses
  21. L22:2728 suppresses
  22. L12:395 suppresses
  23. L19:2011 suppresses
  24. L20:2675 suppresses

Feeds later layers

  1. L26:1028 strongly suppresses
  2. L27:556 strongly suppresses
  3. L25:549 excites
  4. L24:979 suppresses
  5. L27:231 suppresses
  6. L25:572 suppresses
  7. L27:797 excites
  8. L26:2780 suppresses
  9. L27:709 excites
  10. L27:1422 suppresses
  11. L27:180 excites
  12. L27:902 weakly suppresses
  13. L25:1586 weakly excites
  14. L26:484 weakly excites
  15. L27:1454 weakly excites
  16. L24:1406 weakly suppresses
  17. L25:45 weakly suppresses
  18. L27:328 weakly suppresses
  19. L27:663 weakly excites
  20. L25:1187 weakly excites
  21. L26:1102 weakly suppresses fibromyalgia and hour counts
  22. L26:213 weakly suppresses
  23. L26:598 weakly suppresses spot just before a year or century
  24. L25:2924 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 -11.41 and -10.771 values between -8.84 and -8.201 values between -6.92 and -6.271 values between -3.07 and -2.423 values between -2.42 and -1.7819 values between -1.78 and -1.14254 values between -1.14 and -0.502471 values between -0.50 and 0.141078 values between 0.14 and 0.7857 values between 0.78 and 1.438 values between 1.43 and 2.074 values between 2.07 and 2.711 values between 2.71 and 3.351 values between 3.35 and 3.993 values between 3.99 and 4.632 values between 4.63 and 5.281 values between 5.92 and 6.561 values between 7.20 and 7.841 values between 13.62 and 14.26offnegativepositiveoften

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. a
  2. single
  3. single
  4. 单
  5. (single
  6. arded
  7. Single
  8. )a
  9. sing
  10. 單

Suppresses

  1. one
  2. one
  3. One
  4. One
  5. ONE
  6. -one
  7. (one
  8. .one
  9. _one
  10. 一封