Neuron Explorer

Qwen3-0.6B

digit '9' inside years and numbers

A neuron in layer 0 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 a digit inside a year or number, especially the 9 in years like 1993 or 1999. It is mostly about digits in dates and codes. Turning it up gave unrelated text, so the steering does not show a clear 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

    via Getty Images<|endoftext|>Because RNA is such an important part of how the cell builds proteins, some scientists speculated that this was its only job. In 1993, however, Victor Ambros, Rosalind Lee, and Rhonda

  2. #2 peak

    cutting condition can be maintained on the entire path.<|endoftext|>OHIO DEPARTMENT OF AGRICULTURE⏎Division of Plant Industry – Apiculture⏎8995 East Main Street⏎Reynoldsburg, Ohio 4306

  3. #3 peak

    and teachers?⏎According to the Montana Common Core Standards<|endoftext|>Copyright © 1995 Brother Eric Vogel, St. Mary's College⏎Balsamor

  4. #4 peak

    a floral shop.⏎Even more influential of Liberia<|endoftext|>Oxford University Press⏎2017 | 200pp | £7.99⏎Organic chemistry is one of the most exciting areas of modern science, with

  5. #5 peak

    the past 20 years, Florida has been one of the leaders among the 50 states in the standards and accountability movement. In 1999, continuing in this tradition, the Florida Legislature adopted the A+ Plan for Education

  6. #6 peak

    , your<|endoftext|>Longwall dust control-where we are and where we need to go in the 21st century.⏎Longwall USA, 1998 Jun; :123-138⏎Although the number 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: I have been thinking about lately. I have been thinking about lately. I have been thinking about lately. I have been thinking about lately. I have been thinking about lately. I have been thinking about lately. 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

    CT requires strong<|endoftext|>90 LettersLanguage brain teasers are those that involve the English language. You need to

  2. #2 peak

    School of Mathematics Doctoral Training Grant funding - please contact us for more information.<|endoftext|>História (São Paulo)⏎On-line version ISSN 1980-4369⏎DOMINGUES, Beatriz Helena. The guar

  3. #3 peak

    scholar Jean O’Brien: Drop Dead<|endoftext|>Note: Supplemental materials are not guaranteed with Rental or Used book purchases.⏎- ISBN: 9781908049490 | 1908049

  4. #4 peak

    $800 million dollars a day-$13,000 a second-to spend

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

None: this is the first layer.

Feeds later layers

  1. L4:2938 excites
  2. L6:30 excites
  3. L4:308 weakly excites
  4. L1:1148 weakly suppresses
  5. L16:353 weakly excites
  6. L1:18 weakly suppresses
  7. L2:1594 weakly excites the number '100' in numerals
  8. L7:1793 weakly suppresses
  9. L18:196 weakly excites
  10. L2:400 weakly excites
  11. L26:887 weakly excites
  12. L6:2631 weakly excites
  13. L3:2880 weakly excites
  14. L2:2449 weakly excites
  15. L13:2655 weakly excites
  16. L22:2675 weakly excites
  17. L23:2465 weakly excites
  18. L16:917 weakly suppresses 'of' inside organization names
  19. L10:2038 weakly excites
  20. L10:757 weakly suppresses
  21. L26:2647 weakly suppresses
  22. L1:5 weakly suppresses
  23. L12:1747 weakly suppresses
  24. L5:332 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.

3 values between -1.99 and -1.871 values between -1.87 and -1.742 values between -1.62 and -1.501 values between -1.01 and -0.892 values between -0.89 and -0.773 values between -0.64 and -0.523 values between -0.52 and -0.4012 values between -0.40 and -0.2849 values between -0.28 and -0.16866 values between -0.16 and -0.032760 values between -0.03 and 0.09184 values between 0.09 and 0.2110 values between 0.21 and 0.332 values between 0.33 and 0.461 values between 0.58 and 0.701 values between 0.82 and 0.942 values between 1.07 and 1.192 values between 1.31 and 1.431 values between 2.04 and 2.163 values between 2.77 and 2.90offnegativepositiveoften

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. WithValue
  2. ECH
  3. atural
  4. aments
  5. ETA
  6. .toHexString
  7. ),
  8. criptors
  9. ograf
  10. 氾

Suppresses

  1. allis
  2. ้ง
  3. yne
  4. link
  5. null
  6. uro
  7. core
  8. lip
  9. Stations
  10. rophic