Neuron Explorer

Qwen3-0.6B

'of' inside organisation names

A neuron in layer 15 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 'of' when it sits inside the name of an institution or event, such as 'Royal Society of Literature' or 'Fetal Medicine Foundation of London'. Turning it up produced off-topic, rambling text, so the steering does not show a clear effect.

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 plan doesn’t answer is how the city will cope with the climate change that has already been set in motion.⏎Enter a study prepared by Oakland’s Pacific Institute for the California Energy Commission, published in July but not widely circulated until this month.

  2. #2 peak

    Orionid meteor shower. Unlike the displays of previous years, however, the Orionids of 2012 have not put on a very good show. Sky watchers

  3. #3 peak

    . They have been reintroduced to National Reserves of Mongolia and China⏎CONSERVATION STATUS: Classified as Extinct in the

  4. #4 peak

    cooperate with each other is usually used to distinguish humans from other species.⏎“The dogs actually behave benevolently towards other dogs,” said lead researcher Friederike Range of the Messerli

  5. #5 peak

    private doctor⏎(The National Organization for Health Care Services)⏎Advanced Imaging Technology⏎- Ultrasound nuchal translucency (Certified by Fetal Medicine Foundation of London)⏎In humans, there are 23 types of chromosomes and most people

  6. #6 peak

    San Francisco⏎On November 10th, Frida and Diego arrive in San Francisco where Diego has been commissioned to paint murals in the Luncheon Club of the San Francisco Stock Exchange and the California School of Fine Arts [now the San

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 and others like my own ideas, my ideas, and my ideas as much as what and what about what it matters, and what it matters, what it is something to be and what it is, something to be and what

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

    ichia coli (E. coli), and against Streptococcus canis, which can be passed from companion animals to humans.⏎What diseases can you get from dog saliva?⏎This review focused on the most important viral and bacterial zoonotic

  2. #2 peak

    does not deserve all the blame. States such as California, Kentucky and New York began implementing zero tolerance policies for drugs, fighting and gang activity as early as 1989. By

  3. #3 peak

    reading, math, and writing. There are many resources available for parents who want to help their children brush up on their academic skills, including work

  4. #4 peak

    Tina explains how we can study fossilized feces and we reminisce about Jurassic Park.⏎Then, we answer fan-submitted Cosmic Queries. Can we learn anything from early humans that might help us adapt to climate change in the future? We explore

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. L12:234 excites
  2. L4:4 suppresses
  3. L10:132 excites
  4. L13:132 excites "of" after a noun
  5. L10:469 suppresses 'of' after a noun phrase
  6. L11:153 suppresses
  7. L14:917 excites
  8. L14:1519 excites
  9. L6:29 weakly excites
  10. L9:1358 weakly excites
  11. L11:577 weakly excites
  12. L14:747 weakly excites
  13. L8:11 weakly excites
  14. L6:2907 weakly suppresses
  15. L11:558 weakly excites
  16. L14:453 weakly suppresses
  17. L5:763 weakly suppresses
  18. L9:1025 weakly excites
  19. L12:2499 weakly excites
  20. L6:388 weakly excites
  21. L10:2104 weakly excites opening parenthesis in formulas and citations
  22. L6:2111 weakly excites "that" starting a relative clause
  23. L9:463 weakly suppresses
  24. L5:259 weakly excites

Feeds later layers

  1. L16:2834 suppresses
  2. L16:1695 weakly excites
  3. L16:474 weakly excites
  4. L17:1111 weakly excites
  5. L17:1717 weakly suppresses
  6. L22:852 weakly suppresses
  7. L25:334 weakly excites
  8. L23:122 weakly excites
  9. L17:1161 weakly excites
  10. L17:2179 weakly excites
  11. L25:1135 weakly suppresses
  12. L20:756 weakly excites
  13. L20:1672 weakly suppresses
  14. L16:662 weakly excites
  15. L20:277 weakly excites
  16. L16:314 weakly suppresses
  17. L17:76 weakly suppresses
  18. L20:242 weakly excites
  19. L20:1303 weakly excites
  20. L16:2005 weakly suppresses
  21. L21:243 weakly excites
  22. L26:1167 weakly excites
  23. L19:294 weakly excites
  24. L18:2045 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 -5.76 and -5.521 values between -5.05 and -4.811 values between -4.34 and -4.101 values between -4.10 and -3.861 values between -3.86 and -3.633 values between -3.63 and -3.392 values between -3.39 and -3.152 values between -3.15 and -2.9110 values between -2.91 and -2.683 values between -2.68 and -2.4414 values between -2.44 and -2.2011 values between -2.20 and -1.9720 values between -1.97 and -1.7325 values between -1.73 and -1.4928 values between -1.49 and -1.2546 values between -1.25 and -1.0270 values between -1.02 and -0.78103 values between -0.78 and -0.54189 values between -0.54 and -0.30658 values between -0.30 and -0.071836 values between -0.07 and 0.17483 values between 0.17 and 0.41204 values between 0.41 and 0.6584 values between 0.65 and 0.8839 values between 0.88 and 1.1224 values between 1.12 and 1.3617 values between 1.36 and 1.6014 values between 1.60 and 1.837 values between 1.83 and 2.075 values between 2.07 and 2.313 values between 2.31 and 2.541 values between 3.02 and 3.261 values between 3.26 and 3.491 values between 3.49 and 3.73offnegativepositiveoften

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. ernen
  2. McKay
  3. ,err
  4. ��
  5. wyjątk
  6. ünd
  7. ansen
  8. setMessage
  9. <data
  10. ern

Suppresses

  1. agents
  2. prov
  3. cot
  4. 其所
  5. 工作者
  6. 릴
  7. pro
  8. tas
  9. riers
  10. sil