Neuron Explorer

Qwen3-0.6B

gap before a second number

A neuron in layer 11 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 empty spot right before a second number in a pair or range, such as the "12" in "7,12" or the "1903" after "and". Turning it up gave replies that did not clearly relate to numbers, so the steering was not informative.

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

    :3-6)⏎- How did Herod scheme against Jesus? How did God protect Jesus from Herod? (Matthew 2:7-8, 12)⏎- Where did the wise men find the star? (Matthew

  2. #2 peak

    States national anthem, "The Star Spangled Banner"—but was commonly known as 'Scott'. He spent 1898–1901 and 1903–1908 in Buffalo, New York, where

  3. #3 peak

    as Demonstrated by Rubidium-Marked Eggs," American Journal of Tropical Medicine Hygiene⏎52 (2):177-179, 1995).⏎View additional Information on Aedes Aegypti

  4. #4 peak

    zua was described by Marcus (1955) from the coast of Sao Paulo, Brazil. 3 specimens in June 1953, 20 eating Codium in April 1954. Intertidal

  5. #5 peak

    adults get at least 2 1/2 hours of moderate-intensity cardio exercise each week. Alternately, you can perform an hour and 15 minutes of vigorous cardio exercise. By regularly performing cardio exercise, you can

  6. #6 peak

    of the plate boundary types, which type of boundary has the most earthquakes on this list?⏎Put the settings on 7 days, 4.5+ worldwide. Find a region where there is a high concentration of

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 number of 3D objects. I think about the number of points and objects. Maybe I have a lot of points or 1000,000 objects, and I'm thinking about 3D objects. 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

    3(3), pages 315-320.⏎- Angus Deaton & Olivier Dupriez, 2

  2. #2 peak

    lorescence: raceme; bracts leaf-like⏎Flower generally nodding, bell- or cup-shaped; perianth segments 6, of 2 similar whor

  3. #3 peak

    after birth.⏎Pig clones have popped up in the United States, the United Kingdom, Japan and other locations,<|endoftext|>Cleopatra, queen of Egypt and lover of Julius Caesar and Mark Antony, takes her life following the defeat of her

  4. #4 peak

    not able to check, treat for low blood sugar anyway.⏎Give your child 15 to 20 grams of fast-acting sugar such as 3 to 4 glucose tablets, a glass of nonfat milk, or 4 ounces

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. L7:25 suppresses
  2. L6:2945 suppresses
  3. L8:70 excites 'and' between stacked superlatives
  4. L9:1449 suppresses
  5. L6:1084 suppresses
  6. L6:427 suppresses
  7. L6:1706 suppresses
  8. L9:2010 suppresses
  9. L9:212 excites
  10. L6:172 excites
  11. L10:2944 excites
  12. L7:576 suppresses
  13. L8:593 suppresses
  14. L10:2046 suppresses
  15. L9:229 excites
  16. L9:92 suppresses "and" or "or" joining two items
  17. L10:1171 excites
  18. L9:2726 excites
  19. L9:1374 excites
  20. L8:57 suppresses
  21. L4:57 suppresses
  22. L3:250 excites
  23. L9:2516 excites
  24. L8:1608 excites

Feeds later layers

  1. L12:1212 excites
  2. L12:1197 excites
  3. L12:959 excites
  4. L14:844 excites
  5. L12:123 suppresses
  6. L15:38 suppresses
  7. L16:492 suppresses
  8. L16:1294 excites
  9. L14:982 excites
  10. L12:1756 excites
  11. L12:710 excites
  12. L13:517 excites comma closing a list item
  13. L13:125 excites
  14. L15:217 suppresses
  15. L17:84 suppresses
  16. L15:986 excites
  17. L13:371 suppresses
  18. L21:735 excites 'or' in ranges and alternatives
  19. L13:537 suppresses
  20. L13:716 excites
  21. L14:745 excites
  22. L14:2054 excites
  23. L15:745 excites
  24. L15:1162 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 -4.48 and -4.272 values between -3.41 and -3.192 values between -3.19 and -2.981 values between -2.54 and -2.332 values between -2.33 and -2.117 values between -2.11 and -1.908 values between -1.90 and -1.687 values between -1.68 and -1.4712 values between -1.47 and -1.2514 values between -1.25 and -1.0313 values between -1.03 and -0.8237 values between -0.82 and -0.60162 values between -0.60 and -0.39661 values between -0.39 and -0.172110 values between -0.17 and 0.04618 values between 0.04 and 0.26106 values between 0.26 and 0.4852 values between 0.48 and 0.6919 values between 0.69 and 0.9120 values between 0.91 and 1.1213 values between 1.12 and 1.347 values between 1.34 and 1.5513 values between 1.55 and 1.774 values between 1.77 and 1.987 values between 1.98 and 2.203 values between 2.20 and 2.421 values between 2.42 and 2.632 values between 2.63 and 2.852 values between 3.49 and 3.712 values between 3.93 and 4.14offnegativepositiveoften

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. thed
  2. bite
  3. etic
  4. naments
  5. 问他
  6. 벤
  7. vale
  8. ections
  9. HECK
  10. .getB

Suppresses

  1. .TypeString
  2. 扩大
  3. ymb
  4. 这点
  5. oli
  6. 这方面
  7. 这一点
  8. 开头
  9. Ini
  10. Reject