Neuron Explorer

Qwen3-0.6B

acronyms and 'case' in running text

A neuron in layer 3 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 responds to a few scattered words: acronyms like 'OAS' and 'MDC', and the word 'case' in phrases like 'the case'. The examples do not share a clear theme, so it is hard to say what unifies them.

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

    0 students, and has been expanded to include Argentina and Brazil with the support of the Organisation of American States (OAS).⏎The Organic Gardens programme recommends the use of greenhouse-style protected gardens so that children

  2. #2 peak

    a dog’s leg and break the ankle of a human, and the poison causes gangrene if untreated. They live along the banks of rivers and oasises.⏎They ambush creatures, hiding in the sand and snapping when a foot or

  3. #3 peak

    debt to women in dictatorships.”⏎Now, the CIDH has brought the ESMA exhibit to the Organization of American States (OAS) Headquarters in Washington DC in an attempt to start a conversation about crimes against women

  4. #4 peak

    and pancreatic cancer.⏎A family history of breast cancer, male or female, increases the man’s chance of inheriting the disease, though this isn’t always the case — but as with women, many men have no obvious

  5. #5 peak

    they just talk past each other. An additional challenge for a mediator, on top of getting the parties to listen to one another, is that it’s often the case that a person

  6. #6 peak

    have made sure that they left the mark on the city be designing different pieces of beautiful architecture that are itself an amazing piece of work.⏎Delhi will drive anyone crazy when get a views of its age old style and zealous past. 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: that is, a problem of their own. So, the problem is to find out how much time the problem has taken. So, the problem is to find out how much time the problem has taken. So, the problem is to find

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

    INO - ARMOUR - ARTWORK - BIOLOGY - BLACK BOX - COMPUTERS - ELECTRONICS - ENERGY - FRAME - HEAD - JAWS - JIMMY WATSON - KITS - LEGS - MECHANICS - MOTORS

  2. #2 peak

    ?⏎- What are high fidelity prototypes?⏎- What are the three key points in a relationship framework?⏎- Why is it called LOFI?⏎- What does Lo Fi mean?⏎- What is a wireframe diagram?⏎- When should I use low

  3. #3 peak

    , and more than one story is recorded of people succumbing to hunger and disease while still in the harbor.⏎The remarkable confluence of four men made the Jeanie Johnston a exception to this rule. In the New World, there was John M

  4. #4 peak

    writing of the user⏎instructions. Completed CAD models are also used in the wider⏎organization, such as for marketing and packaging.⏎source : wikipedia⏎How lego used to do it⏎do you know?⏎How people do it themselves⏎L

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. L1:2336 weakly excites
  2. L2:1321 weakly suppresses
  3. L0:2814 weakly excites
  4. L2:1012 weakly excites
  5. L1:1574 weakly suppresses
  6. L2:1064 weakly suppresses
  7. L0:1038 weakly excites
  8. L0:741 weakly suppresses
  9. L1:128 weakly excites
  10. L2:111 weakly suppresses
  11. L0:797 weakly suppresses
  12. L1:1154 weakly excites
  13. L0:1301 weakly suppresses
  14. L2:2790 weakly suppresses
  15. L1:491 weakly excites
  16. L1:2358 weakly suppresses
  17. L2:2271 weakly suppresses 'in' in 'take in' phrases
  18. L0:552 weakly excites
  19. L1:2643 weakly suppresses
  20. L2:1494 weakly suppresses
  21. L2:1790 weakly excites
  22. L1:1327 weakly suppresses
  23. L1:571 weakly excites
  24. L1:2176 weakly excites

Feeds later layers

  1. L4:146 excites
  2. L8:2783 weakly excites
  3. L5:2799 weakly excites
  4. L23:1227 weakly suppresses
  5. L13:473 weakly excites
  6. L17:2555 weakly excites
  7. L20:263 weakly suppresses
  8. L13:1987 weakly suppresses
  9. L19:2337 weakly excites
  10. L9:1187 weakly excites
  11. L4:991 weakly suppresses
  12. L12:2126 weakly suppresses
  13. L5:90 weakly excites
  14. L5:2456 weakly excites
  15. L21:1358 weakly suppresses
  16. L12:1338 weakly suppresses
  17. L15:2189 weakly suppresses
  18. L10:2857 weakly suppresses
  19. L9:2024 weakly excites
  20. L27:2990 weakly suppresses
  21. L6:2612 weakly suppresses
  22. L18:2937 weakly excites
  23. L19:2303 weakly suppresses
  24. L19:223 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 -0.90 and -0.831 values between -0.83 and -0.771 values between -0.59 and -0.522 values between -0.52 and -0.462 values between -0.46 and -0.408 values between -0.40 and -0.3433 values between -0.34 and -0.2893 values between -0.28 and -0.21253 values between -0.21 and -0.15424 values between -0.15 and -0.09788 values between -0.09 and -0.031129 values between -0.03 and 0.03648 values between 0.03 and 0.10302 values between 0.10 and 0.16126 values between 0.16 and 0.2250 values between 0.22 and 0.2827 values between 0.28 and 0.346 values between 0.34 and 0.415 values between 0.41 and 0.472 values between 0.47 and 0.533 values between 0.53 and 0.591 values between 0.84 and 0.901 values between 0.90 and 0.971 values between 0.97 and 1.031 values between 1.52 and 1.59offnegativepositiveoften

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. ận
  2. afe
  3. active
  4. embre
  5. isse
  6. lder
  7. eed
  8. צו
  9. lide
  10. olar

Suppresses

  1. ;',
  2. inds
  3. isposable
  4. キ
  5. strings
  6. (savedInstanceState
  7. 膑
  8. InstanceState
  9. ail
  10. caul