Neuron Explorer

Qwen3-0.6B

'and' joining two parallel subjects

A neuron in layer 14 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 the word 'and' when it links two matching things, like two countries or two groups of people. The strongest negative hits are on commas in numbers and a few unrelated words. Turning it up gave repetitive or unclear text.

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

    Mozambique in 2008 and in Ghana, Nepal and Lao PDR in 2009. The pilots in Lao PDR and Nepal specifically focused on the information and communication needs of Indigenous Peoples and were implemented by

  2. #2 peak

    to seven hours a week for science, technology, history, geography and creative arts; and two to three hours a week for sport and other physical activities.⏎English and maths, which dominate the proposed mandatory outcomes, should make up 45 to

  3. #3 peak

    using a nonlinguistic tone anchoring (adaptation) task and a linguistic selective adaptation task in children and adults with and without dyslexia. Children and adults with dyslexia had elevated tone-frequency discrimination thresholds, but both groups benefited

  4. #4 peak

    the FCAT?⏎The FCAT includes multiple-choice, gridded-response (fill in the blanks) and performance tasks (such as essays). The multiple-choice and gridded-response questions are machine scored. Each performance task test is scored by

  5. #5 peak

    , California has a coalition supporting its statewide GND platform, and Illinois and others have been starting to build their support.⏎Some U.S. cities, like Providence and Portland, have also passed climate

  6. #6 peak

    new footbath is not suitable for treatment of active painful lesions, but as an effective measure to prevent the spread of infection from cow to cow.⏎Sheep and cattle are prone to bacterial hoof infections that can be debilitating. A common condition in

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 problem. I don't know, the problem, the one that we know, the one that, I don't know. the problem, the one that, the one that, the one that, the one that, the 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

    Curtiss and purchase a plane for $7,000 ($157,000 in today’s money). While waiting for it to be built,

  2. #2 peak

    9,490. Hourly rates range from a low of $13.50 to a high of $49.05 with a median hourly wage estimated at $23.80. Most mediators are employed

  3. #3 peak

    generally much weaker than tornadic spouts, and tend to be more short-lived. Tornadic waterspouts start in the cloud layer and extend downwards but fair weather waterspouts start at water level and extend upwards. Fair weather watersp

  4. #4 peak

    more on the type of calories you are consuming, than on the amount of calories you are consuming.⏎2. Choosing only low-fat or diet food⏎Processed, low-fat, or “diet” foods are often considered to be the ideal choice when

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. L13:2451 excites
  2. L12:105 excites
  3. L13:1363 excites
  4. L13:1103 excites
  5. L13:619 suppresses
  6. L13:2908 suppresses
  7. L13:116 suppresses
  8. L13:2662 suppresses
  9. L13:300 suppresses
  10. L11:1251 excites
  11. L11:558 excites
  12. L9:454 excites
  13. L1:1464 suppresses
  14. L8:1175 weakly suppresses
  15. L10:782 weakly suppresses
  16. L13:651 weakly suppresses
  17. L10:3052 weakly suppresses 'the' after 'indicate' or 'is'
  18. L11:644 weakly suppresses
  19. L11:379 weakly suppresses
  20. L13:1562 weakly excites unclear: words that set up a qualifier
  21. L13:2540 weakly suppresses
  22. L12:1514 weakly suppresses
  23. L11:1094 weakly suppresses unclear: scattered web and commerce text
  24. L2:2280 weakly suppresses

Feeds later layers

  1. L15:1209 strongly suppresses
  2. L16:2935 suppresses
  3. L15:2334 suppresses
  4. L15:1492 excites
  5. L15:669 excites
  6. L16:473 suppresses
  7. L15:1067 excites
  8. L15:346 suppresses
  9. L15:752 suppresses
  10. L15:2050 excites
  11. L15:185 excites
  12. L18:2918 suppresses
  13. L15:772 excites
  14. L16:669 suppresses
  15. L15:2071 excites unclear: verbs before an object
  16. L15:1080 suppresses
  17. L15:2328 suppresses
  18. L15:474 excites
  19. L15:1917 suppresses
  20. L15:175 suppresses
  21. L15:1504 weakly suppresses
  22. L15:2229 weakly excites
  23. L16:2329 weakly excites
  24. L15:2333 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.

2 values between -3.08 and -2.911 values between -2.91 and -2.732 values between -2.56 and -2.383 values between -2.38 and -2.217 values between -2.21 and -2.034 values between -2.03 and -1.869 values between -1.86 and -1.6811 values between -1.68 and -1.5124 values between -1.51 and -1.3340 values between -1.33 and -1.1650 values between -1.16 and -0.9888 values between -0.98 and -0.81145 values between -0.81 and -0.63197 values between -0.63 and -0.46312 values between -0.46 and -0.28494 values between -0.28 and -0.11859 values between -0.11 and 0.07619 values between 0.07 and 0.25410 values between 0.25 and 0.42267 values between 0.42 and 0.60145 values between 0.60 and 0.7779 values between 0.77 and 0.9556 values between 0.95 and 1.1236 values between 1.12 and 1.3017 values between 1.30 and 1.4711 values between 1.47 and 1.654 values between 1.65 and 1.824 values between 1.82 and 2.005 values between 2.00 and 2.172 values between 2.17 and 2.353 values between 2.70 and 2.871 values between 2.87 and 3.051 values between 3.75 and 3.92offnegativepositiveoften

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. iaz
  2. Joint
  3. ough
  4. z
  5. зу
  6. 社
  7. Named
  8. ※
  9. 为企业
  10. ordial

Suppresses

  1. theless
  2. adays
  3. ONES
  4. elsif
  5. ethyst
  6. stdClass
  7. остей
  8. ELSE
  9. whales
  10. んだけど