Neuron Explorer

Qwen3-0.6B

empty slot just before a number

A neuron in layer 5 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 at a point where a number is about to appear, such as "all the[[ ]]4 volunteers" or "its[[ ]]47th close pass". The number itself isn't in the text yet. When it was turned up, the model's replies began writing numbers and quantities, like "60% of the day" and "45 years".

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

    . Even if you don't make it all the way to the top of the mountain or hike all 40 miles of the trail, you'll have to hike just to get back

  2. #2 peak

    m) wide. Among all of the 37 shipwrecks discovered at Yenikapi, there were a

  3. #3 peak

    of the giant planet Jupiter and its tiny lava-covered moon Io.⏎When the solar-powered spacecraft completed its 47th close pass (perijove) from Jupiter on December 1

  4. #4 peak

    plane crash survivor to avoid starvation, and abusing a kitten.⏎All the 4 volunteers put the actions on the 'wrong' side of the scale. But

  5. #5 peak

    incorporate it into your daily life, the more you will find you use it.⏎A degree of experimentation is necessary to get exact cooking times for<|endoftext|>Here are the 15 types of distorted thinking:⏎1. Filtering: You take the negative details

  6. #6 peak

    ities, including the arms, legs, hands, and feet.<|endoftext|>From the time of Aristotle (384-322 BC) until the late 1500’s, gravity was

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 I can get more information about the possible. I am not sure if I am getting it. - - - - - - - - - - - - - - - - - - - - - - -

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

    be influenced by his internal characteristics; and the second is in which the possible causes of a person’s actions are related to his situation. This is the theory of attribution in social psychology. However, Allah’s Messenger said,⏎“No child is born except

  2. #2 peak

    some of the dangers of this type of attitude:⏎* If people feel that they have little control over their future, they will be unwilling to put a great deal of effort into maintaining their physical well-being. They will not see the point in exercising

  3. #3 peak

    and disorders, addiction is multifactorial, resulting from a combination of genetic, social, psychological, and environmental forces.⏎Genetics plays an important role in conferring vulnerability to addiction. Genetic factors influence how substances are metabolized and experienced. Such

  4. #4 peak

    idical approaches, which give rise to expectations. Frustrated expectations result in an effective response. Music only has meaning unto itself and the music listener ascribes any extra-musical meaning. This includes any emotional meaning. The unfolding of

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. L1:664 weakly excites
  2. L3:1830 weakly suppresses
  3. L0:2423 weakly excites
  4. L2:2544 weakly suppresses
  5. L3:814 weakly suppresses
  6. L4:53 weakly excites blank slot before a number
  7. L4:1438 weakly excites
  8. L4:292 weakly suppresses
  9. L2:1694 weakly suppresses
  10. L2:2204 weakly excites
  11. L1:636 weakly suppresses
  12. L2:2037 weakly excites the digit 5 in decimals like 1.25
  13. L4:2314 weakly suppresses
  14. L4:654 weakly excites
  15. L1:1237 weakly suppresses
  16. L4:205 weakly excites the digit 2 inside citations
  17. L1:846 weakly suppresses
  18. L1:487 weakly suppresses
  19. L0:1269 weakly suppresses
  20. L2:1922 weakly excites
  21. L4:1456 weakly excites
  22. L2:2192 weakly suppresses
  23. L4:2175 weakly suppresses
  24. L4:299 weakly excites

Feeds later layers

  1. L19:1419 weakly excites
  2. L8:2828 weakly excites
  3. L13:1308 weakly excites
  4. L18:390 weakly excites
  5. L19:1947 weakly excites
  6. L10:868 weakly excites
  7. L26:423 weakly suppresses
  8. L26:172 weakly excites
  9. L8:2749 weakly excites
  10. L15:2091 weakly excites
  11. L25:477 weakly excites
  12. L16:262 weakly suppresses
  13. L21:175 weakly excites
  14. L14:549 weakly excites blank spot right before a number
  15. L27:1036 weakly excites
  16. L16:1717 weakly suppresses
  17. L17:222 weakly excites
  18. L27:628 weakly excites
  19. L10:1081 weakly excites
  20. L14:616 weakly excites
  21. L16:120 weakly suppresses
  22. L16:92 weakly suppresses
  23. L16:64 weakly suppresses
  24. L25:333 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 -1.76 and -1.641 values between -1.09 and -0.983 values between -0.86 and -0.755 values between -0.75 and -0.647 values between -0.64 and -0.5317 values between -0.53 and -0.4223 values between -0.42 and -0.31105 values between -0.31 and -0.20467 values between -0.20 and -0.082010 values between -0.08 and 0.03879 values between 0.03 and 0.14226 values between 0.14 and 0.2572 values between 0.25 and 0.3616 values between 0.36 and 0.4719 values between 0.47 and 0.5911 values between 0.59 and 0.705 values between 0.70 and 0.817 values between 0.81 and 0.926 values between 0.92 and 1.035 values between 1.03 and 1.144 values between 1.14 and 1.252 values between 1.25 and 1.372 values between 1.37 and 1.482 values between 1.59 and 1.702 values between 1.70 and 1.811 values between 1.81 and 1.921 values between 1.92 and 2.033 values between 2.03 and 2.152 values between 2.26 and 2.371 values between 2.37 and 2.481 values between 2.48 and 2.592 values between 2.59 and 2.70offnegativepositiveoften

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. ponsored
  2. ledged
  3. 九十
  4. _ENCOD
  5. lesi
  6. stamp
  7. 提出的
  8. 万名
  9. цикл
  10. 预定

Suppresses

  1. acle
  2. inalg
  3. Freund
  4. nard
  5. keit
  6. Constant
  7. geo
  8. Constants
  9. imore
  10. fug