Neuron Explorer

Qwen3-0.6B

'the' inside journal and institution names

A neuron in layer 14 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 'the' (or 'of') when it sits inside the name of a journal, society or institution, such as 'Proceedings of the National Academy of Sciences'. Turning it up made the model write repetitive, looping text that did not clearly relate to that pattern.

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

    people vigilant and careful,” said psychologist Jack van Honk of Utrecht University in the Netherlands, who led the study published May 24 in* Proceedings of the National Academy of Sciences*.⏎In the study, a few dozen females received half

  2. #2 peak

    , the results obtained may also lead researchers to perform similar studies in more clinically relevant models. Proceedings of the National Academy of Sciences USA (PNAS) has published the study online this week

  3. #3 peak

    cultural heritage, the description rules for this structure originate from this area. The CIDOC CRM data model is used. CIDOC is a section of the International Council of Museums (ICOM), which in turn reports to UNESCO. CRM in turn

  4. #4 peak

    Volcanism Program, 2007. Report on Macdonald (Undersea Features) (Wunderman, R., ed.). Bulletin of the Global Volcanism Network, 32:1. Smithsonian Institution. https

  5. #5 peak

    ated muscle,” published in the Proceedings of the National Academy of Sciences, enlarges a growing body of work linking triclos

  6. #6 peak

    and yet we have little understanding of what determines variation among individuals for voluntary exercise levels," Garland said.⏎Study results appear online Sept. 1 in the Proceedings of the Royal Society B.⏎The researchers began their experiments in 1993 with

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 will not. I will not. I will not. I will not. I will not. I will not... I will not... I will not. I will not... I will not... I will not... I will not.

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

    magnetic quantities; a fourth dimension must be included. The different cgs systems arose from differing choices of the additional dimension. See:⏎The cgs electrostatic system of units⏎The cgs electromagnetic system of units⏎The cgs electric units were

  2. #2 peak

    eyrat. Then, he set up his own business in Grigny, around ten kilometres from Lyon. Such partnership disbanded eleven years later letting Barrelon the only one who has signed his name on stained glasses made since 186

  3. #3 peak

    , extensive research, surveys and finance.⏎The main participants in real estate markets are:⏎Overview of real estate markets⏎- Owner/User – These people are both owners as well as tenants. They purchase houses or commercial property as an investment and also to

  4. #4 peak

    learn for yourself.⏎Converting Fractions - Decimals⏎Learn how to convert fractions to decimals and decimals to fractions. Includes practice problems.⏎Decimal Review - Add - Subtract - Multiply⏎Click on this straight forward review lesson about reading, writing,

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. L10:105 suppresses
  2. L13:132 excites "of" after a noun
  3. L13:1046 suppresses 'in' or 'on' ending a phrase
  4. L9:29 excites opening parenthesis before an acronym
  5. L7:6 suppresses
  6. L11:18 excites
  7. L8:25 suppresses
  8. L11:31 suppresses
  9. L9:2717 suppresses
  10. L10:121 suppresses
  11. L13:212 suppresses
  12. L7:26 suppresses
  13. L5:5 suppresses digit after a label like paper or type
  14. L9:2417 excites
  15. L10:33 suppresses
  16. L9:39 excites
  17. L13:1006 excites
  18. L13:755 suppresses
  19. L8:307 suppresses
  20. L13:1552 excites
  21. L8:11 excites
  22. L9:1298 suppresses
  23. L9:7 weakly excites
  24. L9:270 weakly suppresses

Feeds later layers

  1. L15:2569 excites
  2. L15:376 suppresses
  3. L15:2045 suppresses
  4. L17:1775 excites
  5. L17:375 suppresses
  6. L15:1110 excites
  7. L15:2239 excites
  8. L15:1882 excites
  9. L15:1475 excites
  10. L15:891 suppresses
  11. L15:827 suppresses
  12. L16:1056 excites
  13. L20:1447 excites
  14. L15:229 suppresses
  15. L15:74 suppresses
  16. L15:42 excites
  17. L15:2425 suppresses
  18. L17:186 suppresses
  19. L16:412 excites
  20. L19:155 suppresses
  21. L16:2024 suppresses
  22. L18:467 excites
  23. L15:435 suppresses
  24. L15:1388 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 -3.29 and -3.122 values between -2.95 and -2.781 values between -2.78 and -2.621 values between -2.62 and -2.452 values between -2.45 and -2.282 values between -2.28 and -2.116 values between -2.11 and -1.946 values between -1.94 and -1.7712 values between -1.77 and -1.6115 values between -1.61 and -1.4427 values between -1.44 and -1.2730 values between -1.27 and -1.1043 values between -1.10 and -0.9350 values between -0.93 and -0.77106 values between -0.77 and -0.60177 values between -0.60 and -0.43298 values between -0.43 and -0.26591 values between -0.26 and -0.091086 values between -0.09 and 0.08427 values between 0.08 and 0.24289 values between 0.24 and 0.41225 values between 0.41 and 0.58178 values between 0.58 and 0.75112 values between 0.75 and 0.9282 values between 0.92 and 1.0851 values between 1.08 and 1.2525 values between 1.25 and 1.4225 values between 1.42 and 1.5912 values between 1.59 and 1.7610 values between 1.76 and 1.924 values between 1.92 and 2.092 values between 2.09 and 2.264 values between 2.26 and 2.431 values between 2.43 and 2.601 values between 2.60 and 2.773 values between 2.77 and 2.931 values between 3.27 and 3.44offnegativepositiveoften

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. !),
  2. очные
  3. оч
  4. яти
  5. !).
  6. 致
  7. '}}
  8. PowerPoint
  9. очных
  10. .PLL

Suppresses

  1. uilder
  2. atholic
  3. /root
  4. ansom
  5. ellaneous
  6. 郎
  7. 和其他
  8. ships
  9. construction
  10. Shib