Neuron Explorer

Qwen3-0.6B

the '1' of numbers like 10

A neuron in layer 14 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 1 when it starts the number 10 (or 16, 100) in a numbered list. Turning it up made the text drift and repeat, without a clear connection to counting.

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

    5 Radio⏎- 6 Troubleshooting⏎- 7 Tapping the Wire⏎- 8 Scripting⏎- 9 Tech Support⏎- 10 Bridge Comm. Console⏎- 11 Supplementary Video⏎- The most

  2. #2 peak

    - 7 Are there still Taoists?⏎- 8 Is Taoism a Buddhist?⏎- 9 What are the major beliefs of Taoism?⏎- 10 Why is Taoism the best religion?⏎- 11 What does Tao

  3. #3 peak

    octave" model) on 25, 31.5, 40, 50, 63, 80, 100, 125, 160, 20

  4. #4 peak

    |⏎|7 m||7000 mm|⏎|8 m||8000 mm|⏎|9 m||9000 mm|⏎|10 m||10000 mm|⏎|100 m||

  5. #5 peak

    ; to debate or discuss; as, I reasoned the matter with my friend.⏎9. (v. t.) To support with reasons, as a request.⏎10. (v. t.) To persuade by

  6. #6 peak

    a matrix⏎Find nth maximum⏎Matlab Basics II - Create a vector with a repeated entry⏎Create sequnce 1 4 9 16 25.........⏎Basics: Divide integers to get integer outputs in

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 and in the first and on the and and and and and and and and and and and and and and and and and and and and and and and and and and and and and and and and and and and and and and and and

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

    hardware was often referred to as the 16-bit era, and the previous generation as the 8-bit.⏎About Gaming Consoles⏎1 reply to this topic⏎0 user(s) are

  2. #2 peak

    him from a traffic light post in the streets of the capital.⏎The contents of this article is licensed from www.wikipedia.org under the GNU Free Documentation License. Click here to see the transparent copy and copyright details<|endoftext|>What is Celiac / Coeliac

  3. #3 peak

    a different version of it.⏎Volume (Year): 42 (2011)⏎Issue (Month): 2 (June)⏎|Contact details of provider:|| Web page: http://www.tandfonline.com/VECE

  4. #4 peak

    reference errors to , or , if you are the registered author of the cited work, log in to your RePEc Author Service profile, click on "citations" and make appropriate adjustments.:⏎- Andrew Shepherd, 2001

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. L1:1281 weakly excites
  2. L1:394 weakly suppresses
  3. L13:2631 weakly suppresses
  4. L12:409 weakly suppresses
  5. L11:1574 weakly suppresses
  6. L11:29 weakly excites
  7. L2:591 weakly suppresses
  8. L5:1244 weakly suppresses award and prize names
  9. L1:1973 weakly suppresses
  10. L13:804 weakly suppresses
  11. L1:2696 weakly suppresses
  12. L10:46 weakly suppresses
  13. L6:459 weakly excites
  14. L10:282 weakly suppresses
  15. L8:1462 weakly excites
  16. L13:36 weakly excites
  17. L5:2718 weakly suppresses
  18. L13:1602 weakly excites
  19. L7:22 weakly suppresses
  20. L12:737 weakly suppresses
  21. L2:3006 weakly suppresses
  22. L9:253 weakly excites
  23. L2:2079 weakly excites
  24. L0:74 weakly excites

Feeds later layers

  1. L26:16 excites
  2. L16:2189 suppresses
  3. L15:2168 excites
  4. L18:593 suppresses
  5. L18:949 suppresses
  6. L21:1150 suppresses
  7. L26:288 suppresses
  8. L27:136 weakly suppresses
  9. L16:1868 weakly suppresses
  10. L15:2670 weakly excites
  11. L18:1148 weakly excites
  12. L20:561 weakly suppresses end of document marker
  13. L23:194 weakly excites
  14. L24:303 weakly suppresses
  15. L23:1658 weakly suppresses
  16. L27:445 weakly suppresses
  17. L20:1881 weakly suppresses
  18. L15:2070 weakly suppresses
  19. L16:313 weakly suppresses
  20. L27:2111 weakly excites
  21. L20:166 weakly suppresses
  22. L16:2735 weakly excites
  23. L16:1726 weakly excites
  24. L19:1069 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 -0.78 and -0.752 values between -0.64 and -0.611 values between -0.61 and -0.572 values between -0.57 and -0.542 values between -0.54 and -0.505 values between -0.50 and -0.465 values between -0.46 and -0.433 values between -0.43 and -0.3915 values between -0.39 and -0.3617 values between -0.36 and -0.3229 values between -0.32 and -0.2932 values between -0.29 and -0.2564 values between -0.25 and -0.22114 values between -0.22 and -0.18131 values between -0.18 and -0.14205 values between -0.14 and -0.11280 values between -0.11 and -0.07380 values between -0.07 and -0.04655 values between -0.04 and -0.00677 values between -0.00 and 0.03483 values between 0.03 and 0.07311 values between 0.07 and 0.11181 values between 0.11 and 0.14115 values between 0.14 and 0.1869 values between 0.18 and 0.2146 values between 0.21 and 0.2529 values between 0.25 and 0.2824 values between 0.28 and 0.328 values between 0.32 and 0.354 values between 0.35 and 0.399 values between 0.39 and 0.422 values between 0.42 and 0.463 values between 0.46 and 0.502 values between 0.50 and 0.531 values between 0.53 and 0.571 values between 0.60 and 0.64offnegativepositiveoften

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. ensor
  2. icina
  3. 違って
  4. uzzer
  5. edException
  6. ampions
  7. meille
  8. defends
  9. >Select
  10. gether

Suppresses

  1. !⏎⏎
  2. #⏎⏎
  3. ++⏎⏎
  4. ⏎⏎
  5. .⏎⏎
  6. ',⏎⏎
  7. ×⏎⏎
  8. hab
  9. 。⏎⏎
  10. ')⏎⏎