Neuron Explorer

Qwen3-0.6B

"University" in affiliation names

A neuron in layer 21 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 word "University" when it names a person's workplace, like "professor at the University of Florida". It is mostly quiet on other words, and the steered text didn't show this theme.

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

    Jordan M Prutkin, MD, MHS, FHRS⏎- Associate Professor of Medicine, Division of Cardiology, Electrophysiology Section⏎- University of Washington⏎- Leonard I Ganz, MD, FHRS, FACC

  2. #2 peak

    of Philosophy at the University of Texas at Arlington.⏎—Lawrence Zbikowski, Department of Music, University of Chicago⏎—Anthony Cascardi, University of California, Berkeley⏎—Jesse J. Prinz, John J.

  3. #3 peak

    signs of hoarding disorder.⏎The short answer is: Probably not. Hoarding disorder goes beyond stockpiling in an emergency. I am a psychiatrist at the University of Florida and the director of the Center for OCD, Anxiety and Related Disorders.

  4. #4 peak

    critically ill children when there are no pediatric specialists available in their own communities,” says James Marcin, director of the Children’s Hospital Pediatric Telemedicine Program at the University of California, Davis.⏎“People say a picture is worth a thousand words,”

  5. #5 peak

    familiar?⏎DAVID GROSS (University of California, Santa Barbara): Does it bother us? Absolutely.⏎STEVEN WE

  6. #6 peak

    ’65, Peggy Phelps, Jane Hurley Wilson ’64 and Michael G. Wilson, and the Getty Foundation.⏎Daniela Gonzalez-Pruitt, University of Toronto ’17 and Josephine Ren, SCR ’19⏎Stories

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 if I have the same number to the right and the same number to the rightmost number, and the same number to the rightmost number. So, I have a normal $3$-normal number, which can be considered a

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

    effective border biosecurity and to mitigate and prevent the spread of biological threats throughout the region. However, a 2022 study by the Hanoi University of Public Health and the Global Initiative Against Transnational Organized Crime (GI-TO

  2. #2 peak

    Mining-industry; Underground-mining; Coal-mining; Mine-safety; Ventilation; Coal-dust; Dust-control; Control-technology⏎National Institute for Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh, PA 152

  3. #3 peak

    other experiences, Vladimir says he's determined to carry on spreading the message about HIV.⏎Jaroslav Jedlicka, AIDS Programme Manager at Prague's National Institute of Public Health, says awareness is relatively high in the Czech Republic. However, he

  4. #4 peak

    can learn near optimal policies based on trial and error interaction with the . learning objectives samples. carnegie institute of technology. chemical reaction engineering; feedback control systems. college of fine arts. architectural

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. L10:3060 suppresses
  2. L2:2708 weakly suppresses
  3. L20:1215 weakly suppresses
  4. L11:1994 weakly excites
  5. L17:2075 weakly excites
  6. L4:2677 weakly suppresses
  7. L19:2246 weakly suppresses
  8. L19:212 weakly suppresses
  9. L7:2662 weakly suppresses
  10. L13:2104 weakly excites
  11. L9:903 weakly suppresses
  12. L12:1969 weakly excites
  13. L1:37 weakly suppresses
  14. L9:88 weakly excites
  15. L18:1110 weakly suppresses
  16. L12:1598 weakly suppresses
  17. L15:3043 weakly excites
  18. L20:1254 weakly suppresses
  19. L18:2292 weakly suppresses
  20. L4:2123 weakly excites
  21. L20:1887 weakly suppresses
  22. L6:2855 weakly suppresses
  23. L7:180 weakly excites
  24. L7:99 weakly excites

Feeds later layers

  1. L22:2000 weakly suppresses
  2. L22:1016 weakly suppresses
  3. L22:172 weakly suppresses
  4. L24:2816 weakly excites
  5. L25:2559 weakly excites
  6. L22:1603 weakly excites
  7. L24:1236 weakly suppresses
  8. L24:1757 weakly suppresses
  9. L27:864 weakly excites
  10. L22:2153 weakly excites
  11. L24:2937 weakly suppresses
  12. L24:2426 weakly excites
  13. L27:2815 weakly suppresses
  14. L24:1831 weakly suppresses
  15. L26:1507 weakly excites
  16. L24:697 weakly suppresses
  17. L22:1483 weakly excites
  18. L26:1020 weakly suppresses
  19. L24:2818 weakly suppresses
  20. L22:493 weakly suppresses
  21. L24:1172 weakly suppresses
  22. L22:659 weakly suppresses
  23. L24:454 weakly suppresses
  24. L27:267 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 -19.78 and -18.851 values between -17.92 and -16.991 values between -14.20 and -13.271 values between -13.27 and -12.341 values between -9.56 and -8.633 values between -7.70 and -6.771 values between -5.84 and -4.914 values between -4.91 and -3.982 values between -3.98 and -3.057 values between -3.05 and -2.1214 values between -2.12 and -1.19578 values between -1.19 and -0.263052 values between -0.26 and 0.67121 values between 0.67 and 1.6041 values between 1.60 and 2.5323 values between 2.53 and 3.4611 values between 3.46 and 4.399 values between 4.39 and 5.328 values between 5.32 and 6.257 values between 6.25 and 7.183 values between 7.18 and 8.114 values between 8.11 and 9.044 values between 9.04 and 9.973 values between 9.97 and 10.901 values between 10.90 and 11.832 values between 11.83 and 12.762 values between 12.76 and 13.691 values between 15.55 and 16.482 values between 16.48 and 17.41offnegativepositiveoften

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. →Array
  2. Array
  3. Array
  4. (Array
  5. .ibm
  6. Arrays
  7. ropy
  8. arrays
  9. ARRAY
  10. synagogue

Suppresses

  1. 심
  2. eming
  3. #$
  4. 技术水平
  5. Processes
  6. sammen
  7. _duplicates
  8. 的程度
  9. _processing
  10. /archive