Neuron Explorer

Qwen3-0.6B

'of' in university names

A neuron in layer 16 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 appears in a university name, such as 'University of Arizona' or 'University of Chicago'. Turning it up didn't produce anything about universities, so it does not clearly change what the model writes.

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

    University of Hamburg in Germany, in collaboration with colleagues from the University of Aarhus in Denmark, have synthesized a new superconducting material by growing a

  2. #2 peak

    's Kavli Institute for Particle Astrophysics and Cosmology, in collaboration with Lucio Mayer of the Swiss Federal Institute of Technology in Zurich and the University of Zurich, Chiara Mastropietro of the University of Munich in Germany,

  3. #3 peak

    . Thomas Pichler from the University of Vienna report a method for the bulk production of long acetylenic linear carbon

  4. #4 peak

    in 1972, a BS in Physics from the University of California, Irvine in 1978 and a PhD in Astronomy from the University of Arizona in 1982. His research interests have included the linear and

  5. #5 peak

    and seismologist at the University of Arizona. The researchers suspect that a previously discovered region of mantle downwelling a

  6. #6 peak

    33, where he continued his study of social science and mathematics. Herbert Simon earned both his B.A. and Ph.D. in political science from the University of Chicago.⏎- Full Name⏎- Herbert A Simon⏎- APA Award for D

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: I have been thinking about lately, and I have been thinking about lately, and I have been thinking about lately, and I have been thinking about lately, and I have been thinking about lately, and I have been thinking about lately, 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

    , and then watch the current year Address and see if their expectations were met. In this State of the Union...⏎7th - 12th Social Studies & History⏎President Obama's State of the Union Address 2011⏎

  2. #2 peak

    networks called Alice, Bob, and Eve.⏎Ref: Learning to Protect Communications with Adversarial Neural Cryptography. arXiv - Computer Science > Cryptography and Security (21 October 2016) | arXiv:

  3. #3 peak

    as in a discontinuity analysis on observations near the borders between anglophone and francophone countries.⏎- Colonial rule⏎ASJC Scopus subject areas⏎- Economics and Econometrics⏎- Organizational Behavior and Human Resource Management<|endoftext|>- About AST

  4. #4 peak

    between model and experiment was observed.⏎ASJC Scopus subject areas⏎- Electronic, Optical and Magnetic Materials⏎- Ceramics and Composites⏎- Condensed Matter Physics⏎- Materials Chemistry<|endoftext|>Humans are intimately connected with the physical environment, even

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. L15:2688 excites
  2. L9:218 weakly excites
  3. L15:124 weakly suppresses
  4. L15:474 weakly excites
  5. L4:1059 weakly excites
  6. L14:1755 weakly suppresses
  7. L12:2843 weakly excites
  8. L15:1499 weakly excites
  9. L6:1926 weakly excites
  10. L14:1569 weakly suppresses
  11. L13:2269 weakly suppresses
  12. L15:130 weakly suppresses
  13. L8:2101 weakly excites
  14. L15:1504 weakly suppresses
  15. L15:488 weakly excites
  16. L10:1521 weakly suppresses
  17. L12:345 weakly suppresses
  18. L15:954 weakly excites
  19. L2:2135 weakly excites
  20. L0:1967 weakly suppresses
  21. L15:2041 weakly suppresses
  22. L5:612 weakly suppresses
  23. L15:2637 weakly excites
  24. L15:2279 weakly excites

Feeds later layers

  1. L20:294 suppresses
  2. L17:2264 suppresses
  3. L17:1051 excites
  4. L18:630 suppresses
  5. L17:2962 excites
  6. L26:37 suppresses
  7. L26:732 suppresses
  8. L19:221 suppresses
  9. L18:668 weakly suppresses
  10. L27:89 weakly suppresses
  11. L19:753 weakly suppresses
  12. L17:370 weakly suppresses
  13. L17:1635 weakly excites
  14. L26:70 weakly suppresses
  15. L17:404 weakly excites
  16. L24:226 weakly suppresses
  17. L17:544 weakly suppresses
  18. L26:55 weakly suppresses
  19. L19:299 weakly suppresses
  20. L22:90 weakly suppresses
  21. L19:886 weakly excites the space before a rounded number
  22. L23:21 weakly suppresses
  23. L18:2544 weakly suppresses
  24. L27:2529 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 -1.33 and -1.202 values between -0.96 and -0.843 values between -0.84 and -0.724 values between -0.72 and -0.599 values between -0.59 and -0.4741 values between -0.47 and -0.35144 values between -0.35 and -0.23449 values between -0.23 and -0.111423 values between -0.11 and 0.021086 values between 0.02 and 0.14488 values between 0.14 and 0.26191 values between 0.26 and 0.3849 values between 0.38 and 0.5011 values between 0.50 and 0.632 values between 0.63 and 0.752 values between 0.87 and 0.991 values between 1.48 and 1.601 values between 1.84 and 1.971 values between 3.43 and 3.55offnegativepositiveoften

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. ロン
  2. 的母亲
  3. athe
  4. 晚报
  5. RN
  6. ată
  7. Bain
  8. rin
  9. ativas
  10. mant

Suppresses

  1. ..."
  2. ucing
  3. station
  4. azor
  5. 零
  6. star
  7. ";⏎
  8. ";
  9. istical
  10. omorphic