Neuron Explorer

Qwen3-0.6B

'also' after 'not only ... but'

A neuron in layer 2 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 'also' when it completes a 'not only X but also Y' pattern. Turning it up did not make the model's replies visibly about this pattern.

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

    amount of input from James Cameron.<|endoftext|>4 reasons to eat low GI foods⏎Choosing better carbs not only crowds out empty-calorie, glucose-raising foods but also helps battle many health threats. Here’s what eating more high-fibre, low

  2. #2 peak

    iles, sheep or goats and as adults they like cattle, horses, camels, and some antelope as well. These ticks vector not only Q fever but also heartwater and African tick-bite fever (we’ll meet these later), and

  3. #3 peak

    a or Ayurvedic medicine is holistic, which means, it views the body and mind as a whole. It not only treats a person’s physical complaints but also changes his/her day-to-day or ‘dincharya’ practises to help

  4. #4 peak

    āmi to Kāśī where they spent six years. During this sojourn in the north, the boy not only visited many places of pilgrimage but also picked up

  5. #5 peak

    justice policies.. All of these plans are informed by the Federal GND — they are explicit plans to address not only the climate crisis but also racial justice, the housing crisis, the pandemic and a Just<|endoftext|>Wheels,

  6. #6 peak

    lower risk of heart disease.⏎Scientists have recently discovered that the gene IRS1, called lean gene is not only linked to the risk of diabetes type 2 but also that of heart disease.⏎This gene IRS1 does lower fat, but only sub

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

    select: .⏎You can also select a web site from the following list:⏎Select the China site (in Chinese or English) for best site performance. Other MathWorks country sites are not optimized for visits from your location.⏎Contact your local office<|endoftext|>

  2. #2 peak

    60s, the foundation for plate tectonics was laid at the same time, however, Wegener's groundbreaking theory was turned upside down.⏎Earthquakes are not only terrible natural disasters, they also offer a view inside the Earth

  3. #3 peak

    select: .⏎You can also select a web site from the following list:⏎Select the China site (in Chinese or English) for best site performance. Other MathWorks country sites are not optimized for visits from your location.⏎Contact your local office<|endoftext|>

  4. #4 peak

    living being.⏎Once a promising drug has<|endoftext|>Massive Earthquakes Barely Disturb Earth's Natural Rhythms⏎Yesterday's magnitude-

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. L0:2598 weakly suppresses
  2. L1:507 weakly excites
  3. L1:2079 weakly suppresses
  4. L0:331 weakly excites
  5. L1:2447 weakly suppresses
  6. L1:2793 weakly excites
  7. L0:2956 weakly suppresses
  8. L1:2078 weakly suppresses
  9. L1:1617 weakly suppresses
  10. L0:1019 weakly excites
  11. L0:1707 weakly suppresses
  12. L1:926 weakly suppresses
  13. L0:689 weakly excites
  14. L0:686 weakly excites sentence-starting 'This' at text start
  15. L1:2922 weakly excites
  16. L0:2500 weakly suppresses
  17. L1:2498 weakly excites
  18. L0:2779 weakly suppresses
  19. L0:1689 weakly suppresses
  20. L0:283 weakly excites
  21. L0:2036 weakly excites
  22. L0:1659 weakly suppresses
  23. L1:2906 weakly suppresses
  24. L1:785 weakly suppresses

Feeds later layers

  1. L6:617 excites
  2. L3:460 suppresses
  3. L3:486 excites
  4. L5:1386 suppresses
  5. L5:851 suppresses
  6. L6:207 suppresses
  7. L3:400 excites
  8. L3:308 suppresses
  9. L4:453 suppresses
  10. L3:423 weakly excites
  11. L12:1014 weakly suppresses
  12. L7:1347 weakly excites dash starting a new list item
  13. L7:311 weakly suppresses
  14. L17:729 weakly excites
  15. L11:2540 weakly excites
  16. L16:932 weakly suppresses
  17. L16:558 weakly suppresses
  18. L4:2665 weakly suppresses
  19. L14:271 weakly suppresses
  20. L23:2471 weakly suppresses
  21. L16:2898 weakly suppresses
  22. L16:670 weakly suppresses
  23. L5:462 weakly suppresses
  24. L19:2014 weakly 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 -0.78 and -0.731 values between -0.49 and -0.443 values between -0.44 and -0.391 values between -0.39 and -0.3410 values between -0.34 and -0.2932 values between -0.29 and -0.2554 values between -0.25 and -0.20202 values between -0.20 and -0.15362 values between -0.15 and -0.10584 values between -0.10 and -0.05801 values between -0.05 and -0.01750 values between -0.01 and 0.04486 values between 0.04 and 0.09321 values between 0.09 and 0.14169 values between 0.14 and 0.1970 values between 0.19 and 0.2430 values between 0.24 and 0.2815 values between 0.28 and 0.334 values between 0.33 and 0.383 values between 0.38 and 0.432 values between 0.43 and 0.482 values between 0.48 and 0.531 values between 0.53 and 0.571 values between 0.77 and 0.811 values between 0.81 and 0.861 values between 0.91 and 0.961 values between 1.10 and 1.15offnegativepositiveoften

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. 受到影响
  2. blems
  3. 此处
  4. 受不了
  5. anuts
  6. ように
  7. ksen
  8. sanitized
  9. 分割
  10. separated

Suppresses

  1. arial
  2. ForObject
  3. SessionFactory
  4. ذي
  5. ứ
  6. ách
  7. عرف
  8. eed
  9. _ENGINE
  10. ング