Neuron Explorer

Qwen3-0.6B

'both' before a paired noun phrase

A neuron in layer 21 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 'both' when it introduces a pair, as in 'both primary and secondary' or 'both young and old'. It is quiet, or slightly negative, on words like 'between' and 'from' that come before a pair or a comparison. Steering with it did not produce a clear theme in the sample outputs.

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

    your surroundings, including the weather.⏎Boat under control so you can react to river hazards, obstacles and other boats.⏎Know and be aware of river signs, both natural and man-made, such as holes, wrap rocks, under cuts, rock

  2. #2 peak

    departmental goals, which enable students to:⏎- Know historical facts, narratives, and methods⏎- Evaluate and interpret both primary and secondary sources and construct historical arguments⏎- Communicate effectively orally and in

  3. #3 peak

    all caused by human actions. Without a proper system to manage the waste that we created it would lead to a large scale of contamination in soil and water affecting animals both on land and water with micro-plastics, which in the end causes large scale

  4. #4 peak

    the primary perfect man who exemplifies the morality of God, and see him as their leader and prime spiritual guide.⏎Although the overwhelming majority of Sufis, both pre-modern and modern, were and are adherents

  5. #5 peak

    maximum LST can be mixed. This could be because of already mentioned approximate land type detection. In<|endoftext|>Nutrition plays a very important role in the lives of both young and old. However for seniors, it’s important to be aware of new

  6. #6 peak

    the research process simplifies considerably. Under the former scenario, the first step is to conduct market research: a prospective grower should investigate the current market conditions for both conventional and organically produced fruit. It is a good idea to contact packing houses

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: I want to be a bridge between two other people, and I want to make the most of any of the other two people’s (or any other two) that I come into. I also want to be a bridge between two other people

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

    ) are connected, and a shape is a group of coloured cells that are directly or indirectly connected to each other.⏎The diameter of a shape is the greatest distance between any two of its

  2. #2 peak

    on average. We found an interaction between instruction and home-language status such that English-proficient students from language-minority homes improved more than English-proficient students from English-speaking homes. Limited English-proficiency students, however, did not realize gains

  3. #3 peak

    follows: For a given point, called the focus, and a given line not through the focus, called the directrix, a parabola is the locus of points such that the distance to the focus equals the distance to the directrix.

  4. #4 peak

    Previous Article:||World match.|⏎|Next Article:||Ukraine's Orange Revolution.|<|endoftext|>posted on November 17, 2014 23:23⏎Scientists from the WCS (Wildlife Conservation Society),

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. L20:1052 excites
  2. L19:2423 excites
  3. L17:1914 excites
  4. L19:931 excites
  5. L18:2691 excites
  6. L14:3025 suppresses
  7. L18:486 excites
  8. L16:1013 suppresses
  9. L17:1036 suppresses
  10. L20:872 excites
  11. L20:1856 suppresses
  12. L15:3005 suppresses
  13. L14:3046 excites
  14. L16:1039 excites
  15. L17:2130 suppresses
  16. L18:1866 suppresses
  17. L18:787 suppresses
  18. L9:229 suppresses
  19. L19:1893 suppresses
  20. L15:60 suppresses
  21. L17:898 suppresses
  22. L18:1458 suppresses
  23. L6:1168 suppresses
  24. L14:1763 excites

Feeds later layers

  1. L24:814 excites
  2. L22:1497 excites
  3. L25:943 excites
  4. L25:465 weakly excites
  5. L27:393 weakly excites
  6. L23:762 weakly excites
  7. L26:1466 weakly excites
  8. L24:150 weakly excites
  9. L22:638 weakly excites
  10. L26:2328 weakly excites
  11. L25:2924 weakly suppresses
  12. L22:713 weakly excites
  13. L25:179 weakly suppresses
  14. L26:2612 weakly suppresses
  15. L27:1519 weakly suppresses
  16. L25:2748 weakly excites
  17. L24:2417 weakly suppresses
  18. L23:1674 weakly excites
  19. L27:2157 weakly excites
  20. L27:1089 weakly suppresses
  21. L27:2359 weakly excites
  22. L24:706 weakly suppresses
  23. L24:645 weakly suppresses
  24. L26:2434 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.

19 values between -3.11 and -1.502870 values between -1.50 and 0.11909 values between 0.11 and 1.7156 values between 1.71 and 3.3223 values between 3.32 and 4.929 values between 4.92 and 6.537 values between 6.53 and 8.142 values between 8.14 and 9.743 values between 9.74 and 11.352 values between 12.95 and 14.561 values between 14.56 and 16.163 values between 17.77 and 19.381 values between 22.59 and 24.191 values between 30.62 and 32.222 values between 59.52 and 61.13offnegativepositiveoften

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. fixed
  2. 静态
  3. static
  4. 单
  5. fixed
  6. internal
  7. /raw
  8. single
  9. literal
  10. alone

Suppresses

  1. afterward
  2. Видео
  3. afterwards
  4. (cl
  5. 下一篇
  6. enzie
  7. 最后
  8. 最後
  9. 最后一个
  10. ū