Neuron Explorer

Qwen3-0.6B

opening parenthesis before an acronym

A neuron in layer 24 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 open bracket right after a name, just before its short form appears, like 'Advanced Plasma Power (APP)'. Turning it up made the model drift toward analysis and algorithm talk.

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

    L) — can help predict when athletes should return from sports-related concussions, a multicenter, prospective diagnostic study was launched and led by the Advanced Research Core (ARC) of the Concussion Assessment, Research, and Education (CARE)

  2. #2 peak

    companies that includes Advanced Plasma Power (APP), is building a full-scale gas plant that will prove this technology is commercially

  3. #3 peak

    made in laboratories around the world. And they are among the goals of more than 100 horizon-stretching research projects housed in the Advanced Measurement Laboratory (AML) Complex at the National Institute of Standards and Technology (NIST).⏎Completed

  4. #4 peak

    , it may sound like a pipedream (sorry!) but try telling that to the 20 members of the Young Persons Network who visited Advanced Plasma Power (APP) in Swindon to see for themselves how this dream is becoming a

  5. #5 peak

    The Australian Government is making it easier for new migrants to learn and improve their English language skills by increasing the number of available hours through the Adult Migrant English Program (AMEP).⏎From today, migrants in Australia will

  6. #6 peak

    funding,” Turner said.<|endoftext|>This article touches on cultural, linguistic, and technical issues that arise in video recording stories of Deaf individuals who use American Sign Language (ASL). Whether you are Deaf or hearing, documenting ASL interviews presents

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: if a business is supposed to be a "center of a city," then the center of a city is a center of a center, and a center of a center of a center, and so on, with a cycle of three or more

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

    will make your list alphabetical A to Z. (Descending will reverse the list from Z to A).⏎There is an option of telling the program that your list has a header

  2. #2 peak

    affectionately known as the people’s memorial, until our memorial there had been no truth and reconciliation process for this heinous piece of our shared history. Sankofa Day is the first steps in the truth, healing and reconciliation process done for

  3. #3 peak

    ues to Contaminate Surface Water and Drinking Water in the United States⏎Banned in the European Union and clearly linked to harm to wildlife and potentially to humans, the pesticide atrazine provides little benefit to offset its risks. In 20

  4. #4 peak

    in a variety of different conditions. One notable case of such similarity, adenosine (ADO), shares the purine skeleton and oxygenated pendant groups. Together, the two molecules look like this:⏎Adenosine

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. L21:3022 strongly suppresses
  2. L22:178 strongly suppresses
  3. L22:441 strongly suppresses
  4. L23:2946 excites
  5. L21:915 suppresses
  6. L23:2635 excites
  7. L22:672 suppresses
  8. L22:1545 suppresses
  9. L20:2900 excites
  10. L21:337 excites
  11. L20:2780 suppresses
  12. L21:1046 excites
  13. L22:985 excites
  14. L17:2712 excites
  15. L23:1358 suppresses
  16. L21:1602 excites
  17. L22:2394 suppresses
  18. L21:2651 weakly suppresses
  19. L22:411 weakly suppresses
  20. L20:2808 weakly excites
  21. L21:90 weakly excites
  22. L18:51 weakly excites
  23. L20:1200 weakly suppresses 'and' in 'more and more' doubling
  24. L18:2455 weakly excites

Feeds later layers

  1. L25:568 strongly suppresses
  2. L27:652 strongly suppresses
  3. L27:709 suppresses
  4. L26:411 suppresses
  5. L25:315 excites
  6. L25:1361 suppresses
  7. L26:517 excites
  8. L27:231 excites
  9. L27:173 suppresses
  10. L27:203 weakly suppresses
  11. L26:1164 weakly excites
  12. L26:484 weakly suppresses
  13. L25:1014 weakly suppresses
  14. L27:767 weakly excites
  15. L27:1025 weakly excites
  16. L27:919 weakly suppresses
  17. L27:1042 weakly suppresses
  18. L25:2761 weakly suppresses
  19. L27:1292 weakly excites
  20. L27:548 weakly suppresses
  21. L25:754 weakly excites
  22. L26:2487 weakly excites
  23. L27:618 weakly excites
  24. L26:220 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 -7.29 and -6.782 values between -5.78 and -5.271 values between -5.27 and -4.771 values between -4.77 and -4.273 values between -3.77 and -3.265 values between -3.26 and -2.766 values between -2.76 and -2.2610 values between -2.26 and -1.7520 values between -1.75 and -1.2566 values between -1.25 and -0.75369 values between -0.75 and -0.252555 values between -0.25 and 0.26677 values between 0.26 and 0.7672 values between 0.76 and 1.2641 values between 1.26 and 1.7720 values between 1.77 and 2.2714 values between 2.27 and 2.777 values between 2.77 and 3.278 values between 3.27 and 3.785 values between 3.78 and 4.285 values between 4.28 and 4.784 values between 4.78 and 5.292 values between 5.29 and 5.792 values between 5.79 and 6.292 values between 6.29 and 6.792 values between 6.79 and 7.301 values between 7.30 and 7.801 values between 7.80 and 8.302 values between 9.31 and 9.811 values between 10.31 and 10.822 values between 11.32 and 11.821 values between 12.33 and 12.83offnegativepositiveoften

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. アル
  2. analyzer
  3. (api
  4. articulate
  5. ascii
  6. artic
  7. 닝
  8. alg
  9. 算法
  10. agan

Suppresses

  1. A
  2. A
  3. a
  4. →A
  5. *A
  6. :A
  7. _a
  8. .a
  9. <A
  10. _A