Neuron Explorer

Qwen3-0.6B

'of' inside organization names

A neuron in layer 11 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 lights up on the word 'of' when it sits inside the name of an institution, such as 'Baylor College of Medicine' or 'Ponca Tribe of Oklahoma'. Turning it up or down did not produce a clear change in the model's writing.

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

    correct your footsteps to ensure a good posture and foot health,” adds the 29-year-old, who is an Electrical Engineer from Usha Rama College of Engineering and Technology, Vijayawada.⏎The product was to be launched in

  2. #2 peak

    , 2008 Drift Catcher selected as 2008 Tech Awards Laureate by the Tech Museum of Innovation⏎PAN's air monitoring device, invented by chemist and senior scientist

  3. #3 peak

    September, students at the New Albany club were selected to pilot the program locally.⏎Time Warner Cable — parent company of Insight Communications — partnered with Boys & Girls Club of America and FIRST Robotics to bring the program to 20 chapters across the country

  4. #4 peak

    :⏎Three axes of crystal are⏎of equal length and lie at right angle to each⏎other, e.g. Halite mineral.⏎(ii) Tetragonal System :⏎Three axes make right⏎angles with each other. The two horizontal⏎axes

  5. #5 peak

    was 72.⏎A sister, Casey Camp-Horinek, said her brother had cancer.⏎Mr. Camp, a member of the Ponca Tribe of Oklahoma, was a longtime member of the American Indian Movement, organizing more than

  6. #6 peak

    hinoceros unicorns),⏎- Pigmy Hog (Sus salvanius).⏎Over 426 bird species are reported to be found in Royal Manas National Park:⏎- Great Hornbill (Buceros bicornis),⏎- Ruf

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: and And And And and and and and and and and and and and and and and and and and and and and and

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

    stop the wheel, friction material in the form of brake pads (mounted on a device called a brake caliper) is forced mechanically, hydraulically, pneumatically or electromagnetically against both sides of the disc. Friction causes the disc

  2. #2 peak

    is the study of many proteins at once- or a description of proteins in a cell, tissue or cellular organelle on a global scale.⏎The term Proteomics, and the idea of global analysis stems from Genomics (the first application of

  3. #3 peak

    maps provide information about the existing layering of bedrock and faulting, folding or deformation. They include information about the distribution of rock such as limestone, sandstone, coal and granite. Surficial material maps describe deposits that occur above bedrock

  4. #4 peak

    Silla at its height in 576.⏎(Part of Old Korean)⏎|Religion||Buddhism, Confucianism, Taoism, Korean shamanism|⏎|-||57 BC – 4||H

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. L6:1528 excites
  2. L7:195 suppresses 'of' in set noun phrases
  3. L10:709 suppresses
  4. L5:769 suppresses
  5. L10:305 excites
  6. L5:2513 excites
  7. L5:785 suppresses
  8. L2:1175 suppresses
  9. L7:1759 excites
  10. L10:343 excites
  11. L9:190 excites
  12. L5:699 suppresses
  13. L10:1028 suppresses
  14. L8:395 suppresses
  15. L4:493 excites
  16. L5:2194 suppresses
  17. L5:2225 suppresses
  18. L7:551 weakly suppresses
  19. L3:759 weakly excites
  20. L7:1760 weakly suppresses
  21. L3:1092 weakly suppresses
  22. L1:741 weakly suppresses
  23. L10:1218 weakly suppresses
  24. L7:1086 weakly suppresses

Feeds later layers

  1. L15:1012 strongly suppresses
  2. L15:1250 excites
  3. L14:538 suppresses
  4. L15:798 suppresses
  5. L12:1920 suppresses
  6. L15:2906 suppresses
  7. L16:337 suppresses
  8. L15:587 excites
  9. L12:2828 suppresses
  10. L15:3048 suppresses
  11. L15:904 suppresses
  12. L15:1516 suppresses
  13. L14:1774 suppresses
  14. L12:77 suppresses
  15. L13:74 excites
  16. L16:1590 weakly suppresses
  17. L18:662 weakly suppresses middle pieces of split names
  18. L15:1609 weakly excites
  19. L14:2636 weakly suppresses
  20. L13:35 weakly suppresses
  21. L18:467 weakly suppresses
  22. L15:443 weakly suppresses
  23. L14:342 weakly suppresses
  24. L22:2255 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 -6.48 and -6.291 values between -4.03 and -3.841 values between -3.84 and -3.651 values between -2.90 and -2.712 values between -2.71 and -2.522 values between -2.52 and -2.335 values between -2.33 and -2.142 values between -2.14 and -1.965 values between -1.96 and -1.775 values between -1.77 and -1.582 values between -1.58 and -1.396 values between -1.39 and -1.2014 values between -1.20 and -1.0110 values between -1.01 and -0.8324 values between -0.83 and -0.6428 values between -0.64 and -0.4558 values between -0.45 and -0.26267 values between -0.26 and -0.071483 values between -0.07 and 0.121428 values between 0.12 and 0.31511 values between 0.31 and 0.5045 values between 0.50 and 0.684 values between 0.68 and 0.873 values between 0.87 and 1.06offnegativepositiveoften

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. úa
  2. 사를
  3. UGE
  4. 万美元
  5. ?',⏎
  6. bart
  7. useStyles
  8. PackageName
  9. 岁的
  10. illon

Suppresses

  1. 乖
  2. ations
  3. 墨
  4. gio
  5. YNC
  6. 美
  7. �
  8. vision
  9. phe
  10. чки