Neuron Explorer

Qwen3-0.6B

the blank spot right before a number

A neuron in layer 9 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 space just before a number, usually right after words like 'of', 'by', 'from', or 'increased', where a figure such as a percentage, age, or year is about to appear. Turning it up produced loose, unrelated text, so the steering does not show this theme.

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

    . Fully grown males tend to weigh less than 6 ounces, and females less than 4 ounces. So, how large do chameleons grow? The

  2. #2 peak

    linked to the U.S. explosion in childhood obesity. One study of Massachusetts schoolchildren found that for each additional sweet drink per day, the odds of obesity increased 60 percent.⏎As a result, many schools have stopped selling pop or artificial

  3. #3 peak

    moving along the parabola 4y = (x + 2)^2 in such a way that its x-coordinate is increasing at a constant rate of 2 units per second. How fast is the particle's distance to the point (-

  4. #4 peak

    of heart attacks and 54% of strokes worldwide, the authors write. They note that the risk of heart attack and stroke drops by about half for every 20 mm/Hg reduction in systolic blood pressure (the top number of

  5. #5 peak

    to generation.⏎The biggest risk factor for developing PD is advancing age. The average age for the onset of PD is 60 years. In addition, 50 percent more men are affected than

  6. #6 peak

    U.S. Council of Environmental Quality, the cancer risk to people who drink chlorinated water is 93 percent higher than among those whose water does not contain chlorine. The EPA

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 don't know how to start, or how to know if I'm actually trying to write a story." Okay, let's see. I'm a writer, and this is a common challenge. I need to keep going until 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

    a landscape digital photographer. It does not utilize the camera. The professional photographer takes pictures of natural scenes like beaches, hills, or falls. Modern modern technology has actually made it possible for expert landscape photographers to create pictures

  2. #2 peak

    engaged students into a state of active learning. They are present only physically, with their mind elsewhere. They don’t read the materials, do the homework and don’t ask questions, or engage in discussion. Instead, they just watch or listen

  3. #3 peak

    Table of Contents⏎- 1 How welding works⏎- 2 Which welding is best for welding pipes?⏎- 3 Tools required for pipe welding⏎- 4 Steps to welding a pipe⏎- 5 Welding safety precautions

  4. #4 peak

    ellec 40 for the 4-Bit microprocessors i4004 and i4040, Intellec 8 and Intellec 8 mod 80 for the 8-Bit processors.⏎Due

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. L8:9 excites
  2. L8:187 suppresses
  3. L7:83 suppresses
  4. L8:319 excites
  5. L8:272 excites
  6. L1:193 weakly suppresses
  7. L7:135 weakly suppresses 'of' after 'issue' in citations
  8. L8:361 weakly suppresses
  9. L6:123 weakly suppresses
  10. L3:0 weakly excites
  11. L8:395 weakly suppresses
  12. L1:116 weakly excites
  13. L4:1097 weakly suppresses
  14. L4:541 weakly suppresses
  15. L4:515 weakly excites
  16. L3:757 weakly excites
  17. L5:664 weakly suppresses
  18. L4:348 weakly suppresses
  19. L8:1453 weakly suppresses
  20. L7:184 weakly excites
  21. L8:2161 weakly excites
  22. L2:791 weakly excites
  23. L8:1821 weakly excites
  24. L8:1775 weakly suppresses

Feeds later layers

  1. L10:541 suppresses
  2. L11:135 weakly suppresses
  3. L10:42 weakly suppresses
  4. L10:97 weakly suppresses
  5. L10:62 weakly suppresses
  6. L11:56 weakly suppresses
  7. L13:13 weakly suppresses opening parenthesis after a term
  8. L19:2757 weakly suppresses
  9. L10:407 weakly suppresses
  10. L14:4 weakly suppresses
  11. L12:704 weakly suppresses
  12. L10:1090 weakly excites
  13. L10:497 weakly excites
  14. L11:6 weakly suppresses
  15. L13:1358 weakly excites
  16. L12:887 weakly excites
  17. L15:371 weakly suppresses
  18. L12:869 weakly excites
  19. L11:123 weakly suppresses
  20. L15:3050 weakly excites
  21. L18:2642 weakly suppresses
  22. L27:2673 weakly excites
  23. L24:1037 weakly suppresses
  24. L21:2718 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 -4.82 and -4.632 values between -3.70 and -3.511 values between -3.33 and -3.141 values between -3.14 and -2.951 values between -2.95 and -2.771 values between -2.77 and -2.581 values between -2.58 and -2.393 values between -2.39 and -2.216 values between -2.21 and -2.026 values between -2.02 and -1.846 values between -1.84 and -1.652 values between -1.65 and -1.467 values between -1.46 and -1.2817 values between -1.28 and -1.0926 values between -1.09 and -0.9027 values between -0.90 and -0.7268 values between -0.72 and -0.53116 values between -0.53 and -0.35348 values between -0.35 and -0.161904 values between -0.16 and 0.031023 values between 0.03 and 0.21211 values between 0.21 and 0.4065 values between 0.40 and 0.5928 values between 0.59 and 0.7716 values between 0.77 and 0.969 values between 0.96 and 1.152 values between 1.15 and 1.331 values between 1.33 and 1.523 values between 1.52 and 1.711 values between 1.71 and 1.892 values between 1.89 and 2.081 values between 2.08 and 2.261 values between 2.26 and 2.451 values between 2.45 and 2.64offnegativepositiveoften

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. reds
  2. erez
  3. epad
  4. erness
  5. ermal
  6. eren
  7. 、「
  8. ipline
  9. puts
  10. oplevel

Suppresses

  1. 安东尼
  2. arth
  3. andid
  4. .shared
  5. สรร
  6. ewidth
  7. 堑
  8. @endforeach
  9. 一年一度
  10. =settings