Neuron Explorer

Qwen3-0.6B

unclear: word after 'of', 'one' or 'this' in general prose

A neuron in layer 11 of Qwen3-0.6B. Its strongest snippets are mixed: they peak on different tokens. It is featured, so it has steering outputs.

What this neuron does

This neuron is hard to read. It fires on small words such as 'one', 'this', 'a' or 'from' in general statements about people, groups or time periods, like 'members of one human group'. Its negative side fires on nouns after 'this' or 'the' in technical or academic text. Turning it up did not give a clear theme.

Written by Claude from this neuron's activations and steering outputs. Confidence: low. 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

    business practices and priorities vary by cross-border activities or in simple by nationality. Thus, "Culture is the collective programming of the mind which distinguishes the members of one human group from another". (Geert Hofstede, Culture's Consequences

  2. #2 peak

    later (Association for Women’s Rights and Development).⏎The impact of the financial crisis, however, is worldwide. Of the 1.5 million people living on a dollar or less per day, 70% are women (UN.org).

  3. #3 peak

    set him on a path he has been travelling ever since.⏎“For many decades, I have been talking to audiences in schools, universities, and businesses – you name it – I’ll talk to anybody about climate change. And the first picture I always

  4. #4 peak

    paper on The Age Of Milton Friedman that begins with:⏎The last quarter century has witnessed remarkable progress of mankind. The world’s per capita inflation-adjusted income rose from $5400 in 1980 to $850

  5. #5 peak

    -reaching and prevalent culture that has developed as a result of easier communication is the use of memes. A meme is a conception, technique or custom that escalates from one person to another with the purpose of sending a particular meaning or idea represented by the

  6. #6 peak

    they think that their body cannot store anymore minerals. This is especially common in women with brittle bones who think “Well, I should have consumed more calcium when I was younger…” and then they do not take calcium supplements. These women end up being

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: the first time, and I need to do it. I can't do it. So I need to find out why it's a problem. The person I have been talking about, and I need to do it. So I need to

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

    the remaining habitat may become unsuitable in just 50 years. Populations at the south-western edge of the species’s distribution may be the most viable in the long term.⏎Worthen’s

  2. #2 peak

    כֶנֶת(female) is an agent, so סוכנותmeans agency. This term comes from Biblical Hebrew.⏎You may recognize this root in the word for to understand – לְהָבִ

  3. #3 peak

    ) 2017 Robert Albrecht⏎Copyrights and publishing rights of all the texts on this journal belong to the respective authors without restrictions.⏎This journal is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3

  4. #4 peak

    due to climate change. Assuming that Worthen’s Sparrow has minimal dispersal ability, this analysis points to the populations at the south-western edge of the species’s distribution as the most viable in the long term.⏎BirdLife International (2

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. L9:224 suppresses
  2. L9:676 suppresses
  3. L9:868 excites
  4. L7:1118 suppresses
  5. L10:2418 excites
  6. L7:1626 excites
  7. L9:663 suppresses
  8. L7:1744 excites
  9. L10:264 suppresses
  10. L6:824 suppresses
  11. L4:1059 excites
  12. L6:1528 suppresses
  13. L9:2480 weakly suppresses
  14. L8:2143 weakly suppresses
  15. L8:594 weakly excites
  16. L10:1400 weakly excites
  17. L10:1820 weakly suppresses
  18. L4:972 weakly excites
  19. L7:3039 weakly excites 'of' after a noun phrase head
  20. L7:1490 weakly excites
  21. L8:215 weakly excites
  22. L0:2607 weakly suppresses
  23. L9:1505 weakly excites
  24. L9:188 weakly suppresses 'of' in 'National Institute of' names

Feeds later layers

  1. L12:1421 excites
  2. L13:2601 excites
  3. L12:2349 weakly suppresses
  4. L13:1594 weakly excites
  5. L13:2774 weakly suppresses
  6. L13:1857 weakly excites
  7. L13:2886 weakly excites
  8. L14:2241 weakly excites
  9. L12:670 weakly suppresses
  10. L14:304 weakly suppresses
  11. L13:1386 weakly excites
  12. L12:1797 weakly suppresses
  13. L12:599 weakly suppresses
  14. L14:2282 weakly excites
  15. L14:329 weakly suppresses
  16. L12:1184 weakly suppresses
  17. L14:1760 weakly excites
  18. L13:254 weakly suppresses
  19. L12:921 weakly suppresses
  20. L12:883 weakly suppresses
  21. L12:1170 weakly excites
  22. L13:1119 weakly excites
  23. L17:553 weakly suppresses
  24. L14:972 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 -3.54 and -3.381 values between -2.27 and -2.113 values between -2.11 and -1.952 values between -1.95 and -1.796 values between -1.79 and -1.632 values between -1.63 and -1.483 values between -1.48 and -1.325 values between -1.32 and -1.1615 values between -1.16 and -1.0018 values between -1.00 and -0.8445 values between -0.84 and -0.6852 values between -0.68 and -0.52128 values between -0.52 and -0.37282 values between -0.37 and -0.21565 values between -0.21 and -0.051302 values between -0.05 and 0.11675 values between 0.11 and 0.27347 values between 0.27 and 0.43199 values between 0.43 and 0.59108 values between 0.59 and 0.7556 values between 0.75 and 0.9046 values between 0.90 and 1.0622 values between 1.06 and 1.226 values between 1.22 and 1.385 values between 1.38 and 1.548 values between 1.54 and 1.702 values between 1.86 and 2.011 values between 2.01 and 2.171 values between 2.33 and 2.492 values between 2.65 and 2.81offnegativepositiveoften

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. YNC
  2. uja
  3. untos
  4. entifier
  5. 那边
  6. dậy
  7. fires
  8. ulti
  9. 的职业
  10. uffs

Suppresses

  1. thon
  2. chapter
  3. шла
  4. kowski
  5. path
  6. 才是
  7. truncate
  8. processors
  9. değerlendir
  10. **/⏎⏎