Neuron Explorer

Qwen3-0.6B

"of" after a numbered section or title

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 word "of" when it follows a numbered part of a law, code, or document, such as "Chapter 11 of" or "Title 11 of". Turning it up didn't produce a clear theme in the sample text, so its role beyond that pattern is uncertain.

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

    Constitution Act, 1982. Chapter 11 of the UFA establishes a blueprint for land-use planning processes and requires that consideration be

  2. #2 peak

    , it is the country's duty to discourage parochial, racial, tribal, sectarian, and provincial prejudices among the citizens. Under Article 10A of constitution it is also the state's duty to provide<|endoftext|>The authoritative source of information

  3. #3 peak

    for more information.⏎Whilst the specific process of preparing a formal risk assessment is not a requirement of the CDM regulations, regulation 3(6)(a) of the Management of Health and Safety at Work Regulations does require that risks are controlled in

  4. #4 peak

    Access Iis Home Page?⏎Start Administrative Tools Internet Information Services (IIS) Manager. You can access a website via an IP address by expanding the Sites pane of IIS and selecting the website you want to access. You

  5. #5 peak

    Code or the Arizona Revised Statutes). The statutes that are published in a state’s code are grouped by subject matter into Titles, as in Title 11 of the United States Code (bankruptcy laws). In some states, including California,

  6. #6 peak

    , five Member countries in the Region have enacted comprehensive tobacco control legislation, while others are in the process of formulating the same. As decided by the second session of Conference of Parties held in Bangkok, Thailand in 2

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 title of a poem by the 17th century, the title of the first poem of the first poem of the first, the first, first, first, first, the first, first, first, first, first, first

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

    last two weeks Kuhl had twice been admitted for “exhaustion” and doctors were observing the chronically fatigued young man. Beds of wounded men lined room around him.⏎Everything fell to a disciplined hush when the six-foot,

  2. #2 peak

    offend the masses (or the Temple priests) by placing his portrait on the kingdom’s coins. Instead, Herod, like his Hasmonean predecessors, decorated his coins with depictions of common Jewish religious symbols …<|endoftext|>In the field of

  3. #3 peak

    scent, which can contaminate your food.⏎4. Do not store expired food products⏎Storing products post their “use-by’ dates not only clutters your fridge, but also keeps fresh items out of your reach. Consuming expired

  4. #4 peak

    , the air is very cold, and needs to be heated back up to room temperature before it is released back to the air conditioned area.⏎Relative humidity – the moisture contained in any given air mass relative to the temperature of that air. Homes have

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. L9:2862 strongly suppresses
  2. L8:2970 excites
  3. L12:850 excites
  4. L11:2207 suppresses "of" after a noun in trademark and title phrases
  5. L11:2497 suppresses
  6. L9:2219 suppresses
  7. L12:2680 suppresses
  8. L12:2077 excites
  9. L11:1538 excites
  10. L12:1904 suppresses
  11. L12:2756 excites "of" after a noun of quantity or container
  12. L6:2217 suppresses
  13. L9:2659 excites
  14. L12:1653 suppresses
  15. L11:2860 suppresses
  16. L4:1406 excites
  17. L12:1126 excites
  18. L8:2438 suppresses
  19. L9:2566 excites
  20. L11:437 suppresses
  21. L9:2545 suppresses
  22. L9:1025 excites
  23. L9:1428 excites
  24. L12:2168 excites

Feeds later layers

  1. L16:2852 suppresses 'in' after a noun or participle
  2. L16:680 suppresses
  3. L16:2690 suppresses
  4. L16:3007 suppresses
  5. L17:1574 suppresses
  6. L16:1096 suppresses
  7. L16:2612 suppresses
  8. L17:2267 suppresses
  9. L17:2314 suppresses
  10. L19:947 excites
  11. L19:1345 suppresses
  12. L16:2244 suppresses
  13. L17:2877 suppresses
  14. L15:871 suppresses
  15. L18:2573 suppresses
  16. L17:1522 suppresses
  17. L18:1588 suppresses
  18. L16:1051 suppresses
  19. L19:2363 suppresses
  20. L17:2821 suppresses
  21. L16:1830 excites
  22. L16:2673 suppresses
  23. L19:649 suppresses
  24. L20:1977 suppresses

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 -4.66 and -4.492 values between -2.96 and -2.792 values between -2.62 and -2.452 values between -2.45 and -2.282 values between -2.28 and -2.117 values between -2.11 and -1.943 values between -1.94 and -1.779 values between -1.77 and -1.609 values between -1.60 and -1.4312 values between -1.43 and -1.2612 values between -1.26 and -1.0920 values between -1.09 and -0.9221 values between -0.92 and -0.7528 values between -0.75 and -0.5861 values between -0.58 and -0.41135 values between -0.41 and -0.24438 values between -0.24 and -0.071922 values between -0.07 and 0.10865 values between 0.10 and 0.27268 values between 0.27 and 0.4459 values between 0.44 and 0.6113 values between 0.61 and 0.783 values between 0.78 and 0.953 values between 0.95 and 1.126 values between 1.12 and 1.292 values between 1.29 and 1.461 values between 1.46 and 1.631 values between 1.63 and 1.801 values between 1.97 and 2.14offnegativepositiveoften

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. 相关信息
  2. anh
  3. lopen
  4. 几位
  5. iore
  6. adier
  7. ollen
  8. uhe
  9. cker
  10. -record

Suppresses

  1. brains
  2. minds
  3. utra
  4. nostr
  5. society
  6. 箧
  7. mente
  8. mind
  9. Mind
  10. forall