Neuron Explorer

Qwen3-0.6B

"for" in organisation names

A neuron in layer 23 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 "for" inside the names of organisations, like "National Institute for Play", "International Union for Conservation" or "Agency for Healthcare Research". When turned down, the model got stuck repeating the word "control", so that is the only steering hint and it's weak.

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

    ), the Federal Institute for Drugs and Medical Devices (BfArM) and IFA. 0

  2. #2 peak

    his thoughts on the meaning of happiness, and how he understands spirituality as "contemplative science."⏎Stuart Brown, a physician and director of the National Institute for Play, says that pleasurable, purposeless activity prevents violence and promotes trust,

  3. #3 peak

    61 and 1964 by biologists in Russia. Hence, it is also called as the Russian Red Data Book.⏎Currently, the International Union for Conservation of Nature maintains the Red Data Book. IUCN is the world’s

  4. #4 peak

    nforschung (HIH)⏎CRISPR meets single-cell sequencing in new screening method⏎19.01.2017 | CeMM Forschungszentrum für Molekulare Medizin der Österreichischen Ak

  5. #5 peak

    fruit bat (Styloctenium mindorensis) is found only on the Philippine island of Mindoro. Courtesy of ASU's International Institute for Species Exploration.⏎To date some 1.8 million species have been described,

  6. #6 peak

    ; the study results are based on the widely accepted Greenhouse Gas Protocol [GHGP] established by the World Resources Institute [WRI] and World Business Council for Sustainable Development [WBCSD and were announced during an AMD sponsored media panel on

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: What would control the control control control control control control control control control control control control control control control control control control control control control control control control control control control control control control control control control control control control control control control control control control control control

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

    microbiomes or dysbiosis and increased intestinal permeability as well as other changes. Gut microbiota may contribute to MS pathogenesis by modulating host immunity through the regulation of multiple metabolic pathways.⏎The “Gut Feeling”: Breaking Down the Role

  2. #2 peak

    omyces cerevisiae, are well studied and some of the most important food production micro-organisms. S. cervisiae has been used for centuries for the production of wine and beer as well as bread. But other species, the non

  3. #3 peak

    are just going to explain how hemp differs from marijuana.⏎Hemp has a rich history. It has been grown for many millennia for various purposes, for example, the production of textiles, building materials, food, and skin and hair care products.

  4. #4 peak

    intment, compresses,or alcoholic digestion) for treatment of disorders of the locomotor system and gastrointestinal tract. The leaves and stems have also been used for the treatment of the same disorders

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. L20:585 excites
  2. L20:755 excites
  3. L16:2182 weakly excites
  4. L17:472 weakly excites
  5. L2:2048 weakly suppresses
  6. L19:768 weakly suppresses
  7. L10:1638 weakly excites
  8. L19:769 weakly suppresses
  9. L20:2262 weakly excites
  10. L21:2891 weakly suppresses
  11. L1:386 weakly excites
  12. L19:1608 weakly suppresses
  13. L19:2371 weakly suppresses
  14. L1:2905 weakly excites
  15. L7:92 weakly excites
  16. L9:200 weakly suppresses
  17. L17:1392 weakly suppresses
  18. L17:1488 weakly suppresses
  19. L20:2543 weakly suppresses
  20. L19:655 weakly suppresses
  21. L17:960 weakly excites
  22. L21:239 weakly suppresses
  23. L19:752 weakly suppresses
  24. L11:895 weakly suppresses

Feeds later layers

  1. L27:612 suppresses
  2. L26:346 suppresses
  3. L24:480 suppresses
  4. L27:478 suppresses
  5. L27:1199 suppresses
  6. L24:2515 weakly excites
  7. L24:173 weakly excites
  8. L27:1092 weakly suppresses
  9. L24:747 weakly suppresses
  10. L25:868 weakly excites
  11. L27:666 weakly excites
  12. L24:2218 weakly excites
  13. L25:1353 weakly suppresses "in" before "good" condition
  14. L24:2053 weakly excites
  15. L25:2663 weakly excites
  16. L27:2088 weakly suppresses
  17. L25:793 weakly suppresses
  18. L25:1922 weakly excites
  19. L24:738 weakly suppresses
  20. L25:1512 weakly excites
  21. L26:501 weakly suppresses
  22. L27:1560 weakly suppresses
  23. L27:208 weakly suppresses
  24. L25:443 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.

2 values between -14.95 and -14.402 values between -13.31 and -12.771 values between -12.22 and -11.682 values between -11.68 and -11.143 values between -10.05 and -9.502 values between -8.96 and -8.413 values between -8.41 and -7.874 values between -7.33 and -6.781 values between -6.78 and -6.245 values between -6.24 and -5.694 values between -5.69 and -5.1510 values between -5.15 and -4.607 values between -4.60 and -4.0612 values between -4.06 and -3.5127 values between -3.51 and -2.9722 values between -2.97 and -2.4331 values between -2.43 and -1.8847 values between -1.88 and -1.34112 values between -1.34 and -0.79601 values between -0.79 and -0.252309 values between -0.25 and 0.30509 values between 0.30 and 0.8479 values between 0.84 and 1.3945 values between 1.39 and 1.9322 values between 1.93 and 2.4716 values between 2.47 and 3.0213 values between 3.02 and 3.564 values between 3.56 and 4.116 values between 4.11 and 4.651 values between 4.65 and 5.201 values between 5.20 and 5.742 values between 5.74 and 6.283 values between 6.28 and 6.83offnegativepositiveoften

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. ulas
  2. ضا
  3. /Getty
  4. Comey
  5. <Task
  6. ți
  7. образ
  8. slideUp
  9. orem
  10. 试验区

Suppresses

  1. control
  2. etermination
  3. evaluation
  4. etection
  5. -control
  6. construction
  7. Control
  8. 处理
  9. evaluation
  10. 控制