Neuron Explorer

Qwen3-0.6B

the spot right before a number

A neuron in layer 27 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 position just before a number, such as an age, a year, a dosage, or a count, so the model is about to write digits. Turning it up made the model produce strings of digits, which matches what it seems to be preparing for. Its negative side fires on the digits inside a number.

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

    a young age).⏎Guidelines differ for men at average risk. Some recommend an initial PSA (Prostate Specific Antigen) blood test and clinical examination at age 40, and others recommend starting at age 50. In general,

  2. #2 peak

    with an abiding desire for racial equity.⏎Higher education has never been a colorblind meritocracy. Until the implementation of equal opportunity and affirmative action programs in the 1960s, most colleges and universities were historically white institutions. Schools

  3. #3 peak

    20 were women finally guaranteed the right to vote.”⏎On August 24, 1972, Nixon said:⏎“For the first time in the 195 year history of this country, men and women 18 to

  4. #4 peak

    0, in addition to regular self-exams and exams by your doctor or other health care provider.⏎- Colon and rectal cancer: If you are over age 50, talk to your doctor about getting one or more of several screening tests

  5. #5 peak

    see DESCRIPTION]. If either of these conditions exists, the vaccine should not be administered.⏎Between uses, return the multidose vial to the recommended storage conditions between 2° and

  6. #6 peak

    -Gammon reminds us that while it could be harmful, we should remember that the storm itself could have more dire consequences.⏎"If you're looking at a Category 4 or

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:

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

    samples in one full cycle, you will see that there are 45. Divide 44,100 by 45 and you get 980. (I use 980 hertz here rather than

  2. #2 peak

    water factors:⏎Fat: 9 cal/g Carb: 4 cal/g Protein: 4 cal/g⏎- Percent Daily Values are based on a 2,

  3. #3 peak

    now accepted defeat.⏎The brief, fragmentary Book <|endoftext|>Welfare state in capitalism⏎Since the eruption of the sub-prime mortgage crises of 2007 to 2009, there have been hot debates on how the

  4. #4 peak

    in pyrotechny, imparting an intense red color to flames.⏎- n. A metallic chemical element (symbol Sr) with an atomic number of 38.⏎GNU Webster's 1913⏎- n. (Chem

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. L21:339 strongly excites spot just before a year
  2. L25:146 strongly excites
  3. L26:216 strongly suppresses blank slot before a year or number
  4. L25:854 suppresses
  5. L25:182 excites
  6. L22:2275 excites
  7. L19:581 suppresses blank slot before a number
  8. L24:197 excites
  9. L23:77 excites the space right before a number
  10. L26:121 suppresses
  11. L25:3007 suppresses the zero digit in decimals
  12. L22:743 suppresses
  13. L25:33 excites spot where a number is due
  14. L18:377 excites number right after 'about' or 'less than'
  15. L22:372 excites empty spot before a number or "3D"
  16. L20:962 excites decimal points and dots in numbers
  17. L21:395 excites
  18. L23:199 suppresses
  19. L26:868 suppresses
  20. L4:53 excites blank slot before a number
  21. L26:15 suppresses
  22. L26:103 suppresses blank space before a number
  23. L26:1551 excites
  24. L25:232 suppresses

Feeds later layers

None: this is the last layer.

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.

3 values between -16.05 and -12.553 values between -12.55 and -9.065 values between -9.06 and -5.5763 values between -5.57 and -2.073435 values between -2.07 and 1.42221 values between 1.42 and 4.9191 values between 4.91 and 8.4134 values between 8.41 and 11.9014 values between 11.90 and 15.397 values between 15.39 and 18.892 values between 22.38 and 25.871 values between 25.87 and 29.371 values between 29.37 and 32.861 values between 39.85 and 43.341 values between 50.33 and 53.821 values between 53.82 and 57.311 values between 57.31 and 60.811 values between 64.30 and 67.791 values between 71.29 and 74.781 values between 74.78 and 78.274 values between 78.27 and 81.771 values between 81.77 and 85.263 values between 85.26 and 88.754 values between 88.75 and 92.251 values between 92.25 and 95.744 values between 95.74 and 99.231 values between 102.73 and 106.222 values between 113.21 and 116.701 values between 120.19 and 123.69offnegativepositiveoften

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. 5
  2. 3
  3. 1
  4. 2
  5. 4
  6. 0
  7. 8
  8. 9
  9. 6
  10. 7

Suppresses

  1. ⏎
  2. ⏎
  3. ⏎
  4. bones
  5. ”?
  6. ⏎⏎
  7. ”:
  8. ⏎
  9. '?
  10. ⏎