Neuron Explorer

Qwen3-0.6B

dash at start of list item

A neuron in layer 11 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 fires on a dash that starts a new bullet or list item, usually after a line break. Turning it up produced rambling text with no clear link to lists.

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

    as a facebook live on 15th october download a worksheet here download the october to november schedule in colour download the october to november schedule in.⏎Vocabulary Worksheet Mix Set context pre teach vocabulary these stages go online for

  2. #2 peak

    exposure drifting in from the Port of Oakland.⏎- Parks & Recreation: For the PCA program update in Solano County, stakeholders worked

  3. #3 peak

    00 kg armor piercing bombs, in four sections. 40 B5Ns armed with Type 91 torpedoes, also in four sections).⏎- 2nd Group – targets: Ford Island and Wheeler Field, (55

  4. #4 peak

    , popular unrest combined with dynamic leadership.⏎- The Decree for Land: land would be held communally. Peasants

  5. #5 peak

    ” water color paper cut in half.⏎- Extra fine black Sharpie permanent marker.⏎- Water colors and brushes.⏎- Black

  6. #6 peak

    maps in the volume drawn by Bonne's son. Engraved by Andre. Condition: A nice impression with minor offsetting and a few faint spots.⏎B. Plan du Port Palliser, dans la Terre de Kerguelen

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 question that is being asked... (as it is... the person who is responsible for the analysis, so the conclusion is determined, so the agreement is in place, so it is a no-brainer, so it is a no br

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

    irico⏎| ||Key to the genera|| |⏎|1||Terrestrial shrubs, parasitic upon the roots of other plants||Atkinsonia|⏎|Parasitic shrubs on the branches of trees or shrubs||2|⏎

  2. #2 peak

    Using a bar diagram⏎In the bar diagram, the relationship is clear. This is a simple, one-step problem.⏎650 – 39 = 611 candles sold.⏎Printer’s Ink — 1718⏎

  3. #3 peak

    the beach before he was attacked by a golden retriever.⏎Present Simple: This tense describes an event that is happening at the time of writing or speaking. E.g. I go to the shop to buy some chips.⏎Some other examples include:⏎

  4. #4 peak

    3 : 10101100.00010001.0yyyyyyy.yyyyyyyy => 172.17.0.0/17⏎- We need a network with

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. L3:355 weakly excites
  2. L8:447 weakly suppresses
  3. L6:47 weakly suppresses
  4. L4:517 weakly suppresses
  5. L5:198 weakly suppresses unclear: weak "and" in lists
  6. L7:2102 weakly suppresses
  7. L7:82 weakly excites opening bracket before an acronym
  8. L9:1314 weakly suppresses
  9. L6:1578 weakly suppresses
  10. L10:524 weakly excites
  11. L4:191 weakly excites
  12. L7:968 weakly excites
  13. L4:1687 weakly suppresses
  14. L5:1858 weakly excites
  15. L1:257 weakly excites
  16. L1:87 weakly excites
  17. L9:424 weakly excites
  18. L8:510 weakly suppresses
  19. L3:1535 weakly excites
  20. L5:213 weakly excites
  21. L6:12 weakly suppresses
  22. L1:1 weakly excites
  23. L1:743 weakly suppresses
  24. L6:13 weakly suppresses

Feeds later layers

  1. L17:4 weakly excites
  2. L16:1675 weakly suppresses
  3. L22:44 weakly excites
  4. L12:2256 weakly excites
  5. L12:2520 weakly suppresses
  6. L13:2277 weakly excites
  7. L14:1806 weakly suppresses
  8. L16:3013 weakly suppresses
  9. L15:2662 weakly excites
  10. L13:1982 weakly suppresses
  11. L21:1101 weakly suppresses
  12. L15:1670 weakly excites
  13. L18:3052 weakly suppresses
  14. L13:2166 weakly suppresses
  15. L15:2313 weakly suppresses
  16. L14:558 weakly suppresses
  17. L17:912 weakly suppresses
  18. L15:1716 weakly suppresses
  19. L16:531 weakly excites
  20. L12:483 weakly excites
  21. L16:1762 weakly suppresses
  22. L16:1771 weakly excites
  23. L21:2396 weakly excites
  24. L21:1158 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 -1.92 and -1.811 values between -1.69 and -1.571 values between -1.57 and -1.462 values between -1.46 and -1.345 values between -1.34 and -1.225 values between -1.22 and -1.1110 values between -1.11 and -0.994 values between -0.99 and -0.8718 values between -0.87 and -0.7522 values between -0.75 and -0.6451 values between -0.64 and -0.5286 values between -0.52 and -0.40135 values between -0.40 and -0.29284 values between -0.29 and -0.17481 values between -0.17 and -0.05874 values between -0.05 and 0.06572 values between 0.06 and 0.18412 values between 0.18 and 0.30268 values between 0.30 and 0.42213 values between 0.42 and 0.53155 values between 0.53 and 0.6585 values between 0.65 and 0.7775 values between 0.77 and 0.8844 values between 0.88 and 1.0035 values between 1.00 and 1.1223 values between 1.12 and 1.2318 values between 1.23 and 1.359 values between 1.35 and 1.473 values between 1.47 and 1.583 values between 1.58 and 1.703 values between 1.70 and 1.822 values between 1.82 and 1.945 values between 1.94 and 2.051 values between 2.05 and 2.171 values between 2.29 and 2.401 values between 2.64 and 2.75offnegativepositiveoften

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. ница
  2. javax
  3. ival
  4. nieu
  5. 在一个
  6. coni
  7. earrings
  8. 침
  9. ismatch
  10. 不同类型

Suppresses

  1. reds
  2. arrêt
  3. <i
  4. yer
  5. urma
  6. 口水
  7. ogie
  8. دد
  9. ;</
  10. reece