Neuron Explorer

Qwen3-0.6B

'ic' inside longer words

A neuron in layer 25 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 syllable 'ic' when it sits in the middle of a word or a name, such as 'Avic|enna', 'Kiczenski' or 'longic|rus'. Steering it up made the model fall into repeated 'sum of the sum' loops, and steering it down produced a mix of unrelated text.

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

    carried out research into the kinds of scoliosis treatments available at the time Richard was alive.⏎They included looking at 11th century Persian polymath Avicenna's theories on using traction.⏎Avicenna also advocated the massage techniques pract

  2. #2 peak

    on how you can make Orion the “star” of your nightscape photography. And we’ve added many excellent photos of Orion in the landscape by MaryBeth Kiczenski, Ralf Rohner, and Arnaldo Zannoni. So

  3. #3 peak

    eastern North America becomes British territory except St-Pierre and Miquelon Islands (France). British colonial governments for Quebec, Newfoundland (with Île d'Anticosti and Îles de la Madeleine), Nova Scotia (including present-day N

  4. #4 peak

    tenella. Coralline algae are comparable in abundance to stony corals. Other macroalgae include Halimeda tuna, Dictyota divaricata, Loboph

  5. #5 peak

    imulation (rTMS): Brain activity is triggered through weak electric currents, changing the excitability of corticospinal and corticocortial pathways.<|endoftext|>Doing u-substitution twice

  6. #6 peak

    central to north western Canada. There are four subspecies described by Hall (1981) with two (M.v.interior and M.v. longicrus) occurring in California.⏎This is a common bat found in all the mountain

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 top of a spring spring is on a top weight. A spring is spring under weight. An abocess is an abocess. A spring is spring. A spring is spring. A spring is spring. A spring is spring

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

    plan, and institutional) and key audiences (graduate students, faculty, postdoctoral scholars, administrators, organizations) served.⏎John Locke ‘s work Some Thoughts Regarding Education was written in 1693 and still reflects conventional education priorities in

  2. #2 peak

    . A range of new features have been added to the new edition including: new structure to better meet the needs of teaching qualitative research; a new chapter on Grounded Theory plus updated coverage on the full range of<|endoftext|>Stairs - Did You Have

  3. #3 peak

    as distinguished from the lower cognitive faculties, sense, imagination, and memory, and in contrast to the feelings and desires. Reason comprises conception, judgment, reasoning, and the intuitional faculty. Specifically, it is the intuitional faculty,

  4. #4 peak

    Overton, Reasoning, Logic, and Thought Disorders: Deductive Reasoning and Developmental Psychopathology. D.P. Keating, Structuralism, Deconstruction, Reconstruction: The Limits of Reasoning.⏎".....Of interest to serious

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. L3:698 excites
  2. L4:2456 weakly excites
  3. L24:3048 weakly suppresses
  4. L21:2513 weakly excites
  5. L20:2163 weakly excites
  6. L11:1148 weakly excites
  7. L19:1533 weakly excites "and" inside fixed institution names
  8. L22:2433 weakly suppresses
  9. L20:340 weakly excites
  10. L21:752 weakly suppresses
  11. L9:2482 weakly excites
  12. L17:156 weakly excites
  13. L16:2586 weakly excites
  14. L21:1084 weakly excites
  15. L10:1872 weakly excites
  16. L1:1246 weakly excites "word" after a language origin phrase
  17. L6:1145 weakly suppresses
  18. L22:1702 weakly excites
  19. L7:2317 weakly excites
  20. L4:1103 weakly suppresses
  21. L23:2281 weakly excites
  22. L18:1848 weakly excites
  23. L8:88 weakly excites line breaks inside wrapped sentences
  24. L4:92 weakly excites

Feeds later layers

  1. L26:324 weakly excites
  2. L26:1274 weakly suppresses
  3. L27:1304 weakly excites
  4. L27:1725 weakly excites
  5. L26:1444 weakly suppresses
  6. L27:1778 weakly excites
  7. L26:2702 weakly excites
  8. L26:2296 weakly excites
  9. L27:1286 weakly excites
  10. L27:2990 weakly excites
  11. L26:410 weakly excites
  12. L26:2424 weakly suppresses
  13. L27:2976 weakly excites
  14. L26:472 weakly suppresses
  15. L27:1800 weakly excites
  16. L27:2289 weakly excites
  17. L26:406 weakly excites
  18. L27:1327 weakly excites
  19. L27:2115 weakly excites
  20. L27:219 weakly suppresses
  21. L27:1075 weakly suppresses
  22. L27:814 weakly excites
  23. L27:979 weakly excites
  24. L27:1449 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 -23.25 and -22.051 values between -19.64 and -18.441 values between -16.03 and -14.831 values between -14.83 and -13.631 values between -13.63 and -12.422 values between -11.22 and -10.023 values between -10.02 and -8.817 values between -8.81 and -7.619 values between -7.61 and -6.4115 values between -6.41 and -5.2026 values between -5.20 and -4.0048 values between -4.00 and -2.8087 values between -2.80 and -1.59494 values between -1.59 and -0.392640 values between -0.39 and 0.81290 values between 0.81 and 2.02149 values between 2.02 and 3.2264 values between 3.22 and 4.4223 values between 4.42 and 5.6315 values between 5.63 and 6.8310 values between 6.83 and 8.033 values between 8.03 and 9.235 values between 9.23 and 10.444 values between 10.44 and 11.642 values between 11.64 and 12.843 values between 12.84 and 14.051 values between 14.05 and 15.252 values between 16.45 and 17.661 values between 23.67 and 24.88offnegativepositiveoften

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. ouce
  2. Yue
  3. 思想
  4. quares
  5. dressed
  6. uming
  7. >*/⏎
  8. xfd
  9. uju
  10. kor

Suppresses

  1. =top
  2. 阄
  3. 亿元以上
  4. ab
  5. ович
  6. 公
  7. cen
  8. GetName
  9. she
  10. →ns