Neuron Explorer

Qwen3-0.6B

word-start 'con' fragment

A neuron in layer 26 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 reacts to the 'con' piece at the start of words like 'confluence', 'connotations' or 'conveys'. Its negative side is unclear, and turning it up or down gave loops that do not clearly relate to 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

    an appointment online at www.MyFortMyersDentist.com.<|endoftext|>In a world where 1 in 7 people are displaced, your kindness is just condescension.⏎The pirogues are long and narrow, precision-chise

  2. #2 peak

    and lose meaning or beauty.⏎If you love the poem a whole lot, spending more time and attention on it through memorization just creates additional lovable meanings, connotations, and associations. The poem becomes deeper and richer instead of falling apart.⏎

  3. #3 peak

    ative” (signed “story”) is sometimes preferred over “oral” because of an enduring history of “oral” education that has powerful, often painful, connotations for many Deaf adults. Historically “oral” education meant learning without

  4. #4 peak

    ies, and were soundly defeated in a battle in the western delta.⏎After two years of peace, another, more dangerous coalition, the Sea People, a conglomeration of migrating peoples from Asia Minor and the Mediterranean islands who had previously

  5. #5 peak

    field, was the birthplace of modern Manchester.⏎It was established by 78 AD at a site overlooking the confluence of the rivers Irwell and Medlock.⏎This was a timber fort with

  6. #6 peak

    centered and bathed in sunlight, has become covered in flour while he hurriedly scoops up every last grain from a looted depot. The photograph powerfully conveys the animalistic condition of the city's population in the aftermath of the conflict

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: willververververververververververververververververververververververververververververververververververververververververververververververver

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

    40,000 British, Dutch and German troops 250 miles (400km) in five weeks. This extraordinary feat of strategy and logistics left Marlborough's opponents in disarray. Marlborough's tactics included building

  2. #2 peak

    likes” and Artificial Intelligence (AI) algorithms, machines can accurately predict personality traits such as openness, agreeableness, extraversion, conscientiousness, and neuroticism. He further states that AI algorithms, combined with a small number of

  3. #3 peak

    the journey with Piet, past soldiers and enemies, is fraught with danger. Along the treacherous path to Belgium the three children skate using every bit of speed, courage, and strength they can muster. All the time they

  4. #4 peak

    similar to the ruby-throats but are slightly larger and hardier and are differentiated primarily by their range. The broad-tailed<|endoftext|>A woman of extra-ordinary beauty, born in the Old Bailey. She⏎was twelve years a courtes

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. L17:525 suppresses
  2. L25:1807 weakly suppresses
  3. L18:80 weakly excites
  4. L15:2250 weakly excites
  5. L6:362 weakly excites
  6. L2:2797 weakly suppresses
  7. L24:1553 weakly excites
  8. L2:1808 weakly excites
  9. L5:78 weakly suppresses
  10. L0:1575 weakly suppresses
  11. L16:1890 weakly suppresses
  12. L19:1385 weakly excites
  13. L19:850 weakly suppresses
  14. L7:43 weakly suppresses
  15. L21:1886 weakly suppresses
  16. L17:219 weakly excites
  17. L23:2881 weakly excites
  18. L10:3036 weakly excites
  19. L22:1273 weakly excites
  20. L14:2928 weakly excites
  21. L6:1414 weakly excites
  22. L4:2290 weakly suppresses
  23. L16:84 weakly excites
  24. L3:969 weakly excites

Feeds later layers

  1. L27:1015 weakly suppresses
  2. L27:1768 weakly excites
  3. L27:2750 weakly suppresses
  4. L27:2125 weakly excites
  5. L27:2562 weakly suppresses
  6. L27:1631 weakly suppresses
  7. L27:1539 weakly suppresses
  8. L27:2081 weakly suppresses
  9. L27:2874 weakly suppresses
  10. L27:1280 weakly suppresses
  11. L27:2615 weakly excites
  12. L27:2975 weakly suppresses
  13. L27:2571 weakly excites
  14. L27:674 weakly suppresses
  15. L27:544 weakly suppresses
  16. L27:2303 weakly suppresses
  17. L27:1471 weakly suppresses
  18. L27:2581 weakly suppresses
  19. L27:1749 weakly excites hyphen joining word parts
  20. L27:1834 weakly suppresses
  21. L27:2629 weakly excites
  22. L27:986 weakly suppresses
  23. L27:2469 weakly suppresses
  24. L27:2595 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.

2 values between -33.72 and -32.441 values between -20.91 and -19.632 values between -18.34 and -17.061 values between -17.06 and -15.784 values between -14.50 and -13.224 values between -13.22 and -11.942 values between -11.94 and -10.666 values between -10.66 and -9.388 values between -9.38 and -8.0915 values between -8.09 and -6.8117 values between -6.81 and -5.5331 values between -5.53 and -4.2531 values between -4.25 and -2.9773 values between -2.97 and -1.69614 values between -1.69 and -0.412379 values between -0.41 and 0.88360 values between 0.88 and 2.16150 values between 2.16 and 3.4480 values between 3.44 and 4.7252 values between 4.72 and 6.0026 values between 6.00 and 7.2819 values between 7.28 and 8.569 values between 8.56 and 9.8410 values between 9.84 and 11.132 values between 11.13 and 12.411 values between 12.41 and 13.693 values between 13.69 and 14.973 values between 14.97 and 16.253 values between 16.25 and 17.53offnegativepositiveoften

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. bservable
  2. ично
  3. но
  4. igrations
  5. terior
  6. uppy
  7. pecia
  8. orda
  9. erno
  10. ounds

Suppresses

  1. yen
  2. ')</
  3. 一角
  4. ayar
  5. �
  6. YW
  7. -END
  8. сте
  9. ver
  10. .Pages