Neuron Explorer

Qwen3-0.6B

the 9 in years like 19xx

A neuron in layer 22 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 digit 9 in the middle of a year that starts with 1, such as 1983 or 1948. Its negative side is less clear, mostly showing up on words like 'paying' and 'no' in ordinary sentences. Turning it up did not produce any obvious theme in the output.

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

    . The process put in at home and inexpensive. It will require a bit of specialized knowledge and a few tools.⏎The first computer virus was developed in 1983 by Fred Cohen, a mentor at the School of New Dreamland.

  2. #2 peak

    prove it. Many conspiracy theories regarding secret chambers and passageways within the pyramids exist to this day.⏎These conspiracy theories became even more widespread in 1993, when Dr. Z<|endoftext|>Our cataloguer and book reviewer Jo takes

  3. #3 peak

    within Islam, but the majority in Iran and Iraq. Belief in the messianic Twelfth Imam was a key component of the Iranian Revolution of 1979

  4. #4 peak

    , or whether it’s nitrogen and phosphorus – will make it to the sea. And that means that probably water quality will<|endoftext|>Harlem riot of 1935⏎The Harlem riot of 1935 was Harlem's first

  5. #5 peak

    0 antibody can cure multiple types of cancer and prevent spontaneous genetically driven cancers.<|endoftext|>Do You Know?⏎National Guard Bureau⏎WASHINGTON, March 5, 1998 Our National Parks⏎

  6. #6 peak

    that the UK must reduce the amount of gold in its sovereigns, albeit by<|endoftext|>Oil Crisis and consolidation (1974–1988)⏎The first mild recession in the Federal Republic

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 first part of the course of a class of students. A class of students who will eventually leave a class with the last member, the last class out of the class of students, a class that will have a member who will leave a

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

    score?⏎Credit scores help lenders decide whether to lend money, how much to lend and how much interest to charge. By not knowing their credit score, people risk paying more interest when they take out a loan and may face limits on the amount they

  2. #2 peak

    expression. Digital space offers an affordable, accessible, and, if desired, anonymous, soapbox far freer than its terrestrial counterparts. So moving there is a no-brainer, as natural as water running downhill.⏎But the question that political scientists

  3. #3 peak

    )⏎- Mixed facts practice pages⏎- Answer keys⏎GRAB THE BUNDLE! It will get students EXCITED TO PRACTICE SUBTRACTION! No more boring flashcards over and over. Simply print and students are ready to engage

  4. #4 peak

    eat more healthy meals 75-80% of the time in a week, that’s 16 meals that are healthy. This way there is no feeling of failure if you aren’t eating an entirely healthy meal every time you eat

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. L21:1435 suppresses
  2. L19:1301 excites
  3. L8:2024 weakly excites
  4. L18:2776 weakly excites
  5. L20:919 weakly excites
  6. L19:2792 weakly suppresses
  7. L4:42 weakly excites
  8. L7:71 weakly suppresses
  9. L18:311 weakly suppresses
  10. L15:2101 weakly excites
  11. L7:2895 weakly excites
  12. L5:1177 weakly excites
  13. L8:2458 weakly excites
  14. L10:23 weakly suppresses
  15. L9:345 weakly excites
  16. L15:2189 weakly excites
  17. L14:2519 weakly excites
  18. L12:2401 weakly suppresses
  19. L18:2561 weakly suppresses
  20. L17:292 weakly excites
  21. L8:1990 weakly excites
  22. L13:63 weakly excites
  23. L17:978 weakly suppresses
  24. L15:2754 weakly excites

Feeds later layers

  1. L23:3018 weakly excites
  2. L23:3042 weakly excites
  3. L23:1666 weakly excites
  4. L25:2781 weakly excites
  5. L27:1694 weakly excites
  6. L25:1373 weakly excites
  7. L23:1068 weakly excites
  8. L24:2817 weakly excites
  9. L27:1736 weakly suppresses
  10. L26:1273 weakly excites
  11. L26:2025 weakly suppresses
  12. L23:1622 weakly excites
  13. L26:2351 weakly excites
  14. L27:1480 weakly suppresses
  15. L24:514 weakly excites
  16. L23:978 weakly excites
  17. L24:2017 weakly suppresses
  18. L25:2909 weakly suppresses
  19. L23:787 weakly suppresses
  20. L24:2609 weakly suppresses
  21. L24:2740 weakly excites
  22. L25:445 weakly suppresses
  23. L23:587 weakly excites
  24. L24:2630 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.

2 values between -19.67 and -18.601 values between -16.46 and -15.381 values between -15.38 and -14.313 values between -14.31 and -13.241 values between -13.24 and -12.172 values between -12.17 and -11.092 values between -11.09 and -10.024 values between -10.02 and -8.958 values between -8.95 and -7.8811 values between -7.88 and -6.8113 values between -6.81 and -5.7320 values between -5.73 and -4.6638 values between -4.66 and -3.5954 values between -3.59 and -2.52105 values between -2.52 and -1.44526 values between -1.44 and -0.372622 values between -0.37 and 0.70309 values between 0.70 and 1.7799 values between 1.77 and 2.8540 values between 2.85 and 3.9213 values between 3.92 and 4.9910 values between 4.99 and 6.069 values between 6.06 and 7.145 values between 7.14 and 8.213 values between 8.21 and 9.282 values between 10.35 and 11.423 values between 11.42 and 12.501 values between 13.57 and 14.641 values between 22.15 and 23.22offnegativepositiveoften

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. рест
  2. Wear
  3. Characters
  4. Characters
  5. elan
  6. aged
  7. 唧
  8. итет
  9. 惺
  10. itat

Suppresses

  1. {name
  2. cole
  3. {k
  4. ucket
  5. cole
  6. {[
  7. {n
  8. Cole
  9. MARK
  10. cur