Neuron Explorer

Qwen3-0.6B

the word 'new'

A neuron in layer 8 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 'new' in phrases like 'the new city', 'build new ones' and 'new blood vessels'. Turning it up did not produce any clear theme in the continuations, so the steering shows nothing about this word.

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

    to uphold Ma'at, the principle of order and truth. At death, the Pyramid Texts stated that the pharaoh was to become the sun or the new Osiris. However, the journey of this transformation was perilous. The Pyramid

  2. #2 peak

    specific period⏎of time; at the conclusion of this period, this world will necessarily end and⏎its earth and heavens shall be replaced by a new earth and new heavens; a new⏎world will then come into being. All mankind shall be

  3. #3 peak

    Sometimes it makes sense for a function to return multiple things. In such cases, it's common to just return them in a tuple: "return (count, new_obj)".⏎Sometimes you might want to return objects of different types based on whether

  4. #4 peak

    and deserve full citizenship. They are fighting for their rights and their place in the new scene after the revolution. However their struggle exemplifies the many of the issues that

  5. #5 peak

    28 grams of protein into your diet. This is 13 percent of the daily value for protein, a nutrient your body needs to repair cells and build new ones.⏎Source of Minerals⏎Tamales contain good amounts of iron and phosph

  6. #6 peak

    are carried over to the new data directory.⏎Stop the existing MySQL instance using the data directory. This must be

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 have to be honest with you that I have no time to write for you what I can say. And I also have no time to write for you in the same way. I have been thinking about this the same way I have been

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

    of the earth covered in saltwater, many plants and animals live along, on, and under the ocean. Ocean habitat consists of three zones- the Intertidal shoreline zone, the Pelagic or open ocean zone, and the deep Abyssal

  2. #2 peak

    is generally agreed to be responsible for up to 85% of the visible signs of skin aging! The good news is that, in comparison to intrinsic skin aging, it is

  3. #3 peak

    the signs of skin aging. Now that you<|endoftext|>America's first industry was neither textiles, nor iron,

  4. #4 peak

    of sugary or acidic foods and beverages, and drink plenty of water to help wash away food particles and neutralize the acid in your mouth.⏎5. Floss on a daily basis:⏎Flossing removes food particles and plaque that can become

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. L4:2828 excites
  2. L4:1906 excites
  3. L7:283 suppresses
  4. L3:1615 weakly suppresses
  5. L0:134 weakly suppresses
  6. L7:1753 weakly excites
  7. L2:641 weakly suppresses
  8. L4:2459 weakly excites
  9. L2:572 weakly suppresses
  10. L6:170 weakly suppresses
  11. L7:1521 weakly suppresses
  12. L4:1683 weakly suppresses
  13. L2:2903 weakly suppresses
  14. L5:1310 weakly suppresses
  15. L0:1893 weakly suppresses
  16. L7:2550 weakly excites
  17. L0:1106 weakly excites
  18. L7:2899 weakly suppresses
  19. L6:261 weakly suppresses
  20. L2:1447 weakly excites
  21. L7:700 weakly suppresses
  22. L4:2860 weakly suppresses
  23. L4:1931 weakly suppresses
  24. L5:698 weakly excites

Feeds later layers

  1. L11:1755 excites
  2. L19:1042 excites
  3. L15:1892 excites
  4. L11:1713 excites
  5. L14:2743 excites
  6. L11:1895 excites
  7. L16:2123 excites
  8. L21:2084 excites
  9. L9:409 weakly suppresses
  10. L12:1122 weakly excites
  11. L11:239 weakly suppresses
  12. L9:743 weakly excites
  13. L9:750 weakly excites
  14. L9:803 weakly excites
  15. L12:2917 weakly excites
  16. L16:887 weakly excites
  17. L12:480 weakly suppresses
  18. L17:1620 weakly excites
  19. L12:1854 weakly suppresses
  20. L14:832 weakly excites modifier words before their head word
  21. L13:71 weakly suppresses
  22. L9:1215 weakly excites
  23. L16:2395 weakly excites
  24. L9:1475 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 -1.80 and -1.591 values between -1.17 and -0.974 values between -0.97 and -0.765 values between -0.76 and -0.5548 values between -0.55 and -0.35685 values between -0.35 and -0.142420 values between -0.14 and 0.07658 values between 0.07 and 0.2862 values between 0.28 and 0.489 values between 0.48 and 0.698 values between 0.69 and 0.903 values between 1.11 and 1.311 values between 1.31 and 1.521 values between 1.73 and 1.941 values between 3.18 and 3.391 values between 6.29 and 6.50offnegativepositiveoften

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. replacements
  2. 替代
  3. skirts
  4. ования
  5. 取代
  6. replaces
  7. improves
  8. 重建
  9. versions
  10. improvements

Suppresses

  1. apot
  2. 帻
  3. apest
  4. apat
  5. nuit
  6. çon
  7. Chin
  8. 一杯
  9. flag
  10. acist