Neuron Explorer

Qwen3-0.6B

'and' or 'to' before a second verb

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 responds to the word 'and' (or 'to') that links one action to another, as in 'turn and look' or 'pause and consider'. Turning it up did not produce a visible pattern in the output.

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

    -month-olds.⏎If your baby admires his reflection in a mirror and you suddenly appear behind him, he's likely to turn and look for you instead of believing that you're in the mirror.⏎A simple game

  2. #2 peak

    -old Eric Mohat killed himself. His suicide occurred on the same day a fellow student ridiculed him in class by saying, “Why don’t you go home and shoot yourself, no one will miss you.”⏎Fourteen-year-old Jeremiah Lasater

  3. #3 peak

    ability to bring in subtle elements, to pause and consider is vital for their patient.⏎We now have artificial intelligence which can replicate part

  4. #4 peak

    polygonal transects remains elusive.⏎Keywords: Boundary overlap, Inclusion region, Reflection method, Transect configuration⏎- We recommend that you also print this page and attach it to the printout of the article

  5. #5 peak

    is then delivered to the condenser, where it is cooled, dissipating the heat and steadily converting the gas back to a liquid state. A liquid receiver captures and stores the refrigerant between the condenser and the metering device.⏎Note:

  6. #6 peak

    ospitable area, the Spanish soon inaugurated the practice of sending regular military parties inland to subdue the increasingly hostile natives.⏎Members of the indigenous population were captured and enslaved to support the fledgling colony. The object of Columbus’ desire changed

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 thing that I think is very interesting, and then I have been thinking about it, and then I have been thinking about it, and then I have been thinking about it... and then I have been thinking about it... and then I

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

    on, Drupal stores all the HTML code for any page visited by an anonymous user directly into the database. When another request for the same page comes along, Drupal knows to fetch this page out of the database rather than re-generating it from scratch

  2. #2 peak

    of any kind, including, without limitation, representations and warranties as to the functionality of the translation feature or the accuracy or completeness of the translations.⏎Translations are not retained in our system. Your use of this feature and the translations is subject

  3. #3 peak

    such as cookies and analytics, to customise content, advertising and provide social media features. This will be used to analyse traffic to the website, allowing us to understand visitor preferences and improving our services. Learn more<|endoftext|>Acknowledged as one of

  4. #4 peak

    unable to make enough thyroid hormone have congenital hypothyroidism, meaning they are born with an underactive or absent thyroid gland. If the condition is not found and treated, children can have mental retardation and growth failure. But the good

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. L13:2677 suppresses
  2. L13:2929 excites
  3. L13:433 excites
  4. L12:2384 excites
  5. L13:2398 excites
  6. L12:1812 suppresses
  7. L12:650 weakly suppresses
  8. L13:1523 weakly excites
  9. L12:185 weakly excites
  10. L12:313 weakly suppresses
  11. L12:1279 weakly excites
  12. L8:2810 weakly suppresses
  13. L13:3031 weakly excites
  14. L11:2908 weakly excites
  15. L7:893 weakly suppresses
  16. L12:2207 weakly excites
  17. L11:2006 weakly excites
  18. L13:341 weakly excites
  19. L13:2096 weakly suppresses
  20. L11:814 weakly suppresses
  21. L1:1735 weakly excites
  22. L12:120 weakly suppresses
  23. L12:2439 weakly excites
  24. L5:26 weakly excites

Feeds later layers

  1. L15:273 excites
  2. L16:3060 suppresses
  3. L17:2911 suppresses
  4. L16:2167 suppresses
  5. L17:2495 excites
  6. L16:2550 suppresses
  7. L16:1207 weakly suppresses
  8. L15:1176 weakly excites
  9. L19:2726 weakly suppresses
  10. L18:529 weakly excites
  11. L16:2762 weakly excites
  12. L16:2753 weakly suppresses
  13. L16:2447 weakly suppresses
  14. L19:1991 weakly excites
  15. L16:357 weakly excites
  16. L15:3040 weakly suppresses
  17. L17:3058 weakly excites
  18. L25:119 weakly excites
  19. L16:2420 weakly suppresses
  20. L15:2225 weakly suppresses
  21. L25:245 weakly excites
  22. L15:2363 weakly suppresses
  23. L16:457 weakly suppresses
  24. L16:1687 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 -3.17 and -2.981 values between -2.98 and -2.791 values between -2.41 and -2.221 values between -2.03 and -1.842 values between -1.84 and -1.663 values between -1.66 and -1.474 values between -1.47 and -1.288 values between -1.28 and -1.0912 values between -1.09 and -0.9023 values between -0.90 and -0.7140 values between -0.71 and -0.52110 values between -0.52 and -0.33262 values between -0.33 and -0.141758 values between -0.14 and 0.051226 values between 0.05 and 0.24349 values between 0.24 and 0.4356 values between 0.43 and 0.6220 values between 0.62 and 0.818 values between 0.81 and 1.008 values between 1.00 and 1.193 values between 1.19 and 1.384 values between 1.38 and 1.573 values between 1.57 and 1.761 values between 1.94 and 2.132 values between 2.32 and 2.511 values between 2.51 and 2.701 values between 4.22 and 4.41offnegativepositiveoften

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. aker
  2. umer
  3. Greater
  4. ものは
  5. ount
  6. onas
  7. Enh
  8. ��
  9. noop
  10. SqlParameter

Suppresses

  1. назначен
  2. zeitig
  3. جين
  4. ;!
  5. 下来
  6. icated
  7. ubic
  8. 热度
  9. gio
  10. 学到