Neuron Explorer

Qwen3-0.6B

'turn' in 'which in turn'

A neuron in layer 8 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 'turn' in the phrase 'which in turn', where one thing causes another in a chain. Turning it up pushed the model toward repeating phrases like 'in turn' and 'the temperature of the temperature', so the effect was mostly repetition.

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

    reconfiguration region are governed by field-aligned currents, which in their turn generate auroral brightenings near the foot points of the spacecraft. It is

  2. #2 peak

    , which in turn was divided into the precincts of Beaufort , Hyde , Craven

  3. #3 peak

    production will be the most affected by climate change by 2030. This would lead to food insecurity, which in turn could lead to social instability, as was already the case in Mozambique in

  4. #4 peak

    for bacteria which in turn produce methane gas.⏎Methane gas is a greenhouse gas that traps heat much

  5. #5 peak

    interaction of nature, nurture and the decisions and choices that people make. Biological programming is your genetics. Genes provide the instructions for structuring amino acids which in turn influence the proteins within our body cells thus programming chemical basis of our biology. Our

  6. #6 peak

    anting trees in clearings and along roads. The communities’ low-impact timber harvesting provides jobs as well as resources for health and education. These opportunities in turn have given the communities a long-term stake in protecting the forest.⏎1 // SITE

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 given to the user a... the user has been given to the user's first time, and it's not related to the time or money. A good value is, I would say that it is... and that it is

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 end result is a vertical wall, or dyke,<|endoftext|>-Why is balance important in an ecosystem?⏎-What is the relationship between living and non-living things in an ecosystem?⏎Then I walked them through

  2. #2 peak

    ly described by the Monty Python comedy troupe as “taxing foreigners living abroad.” But the fact that

  3. #3 peak

    ’s, a couple of entrepreneurs leased a designer to type a program who post unwanted ads to help you lots and lots of electronic newsgroups. Exactly as Monty Python’s chanting Vikings drowned aside most other conversation into the the latest eatery,

  4. #4 peak

    is driven by the simple fact that Apache is a hypertext⏎transfer protocol (HTTP) server that runs in the background as a daemon. Because⏎it is a daemon, it must do all the

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. L7:2495 weakly excites
  2. L7:2962 weakly suppresses
  3. L6:2930 weakly suppresses
  4. L7:2472 weakly excites
  5. L7:2721 weakly excites
  6. L7:749 weakly suppresses
  7. L6:326 weakly suppresses
  8. L6:757 weakly excites
  9. L6:2395 weakly suppresses
  10. L6:1747 weakly suppresses technical word fragments, splicing and isotopes
  11. L6:2445 weakly suppresses
  12. L7:1858 weakly excites
  13. L6:2661 weakly excites
  14. L4:1929 weakly suppresses
  15. L2:2703 weakly suppresses
  16. L7:1251 weakly excites
  17. L4:2679 weakly excites
  18. L5:2455 weakly suppresses
  19. L6:2595 weakly suppresses
  20. L7:1559 weakly excites
  21. L4:940 weakly suppresses
  22. L2:704 weakly suppresses
  23. L2:2884 weakly excites
  24. L4:2590 weakly excites

Feeds later layers

  1. L10:769 excites
  2. L10:2126 excites
  3. L16:2338 excites
  4. L12:2380 excites
  5. L15:1741 excites
  6. L14:2295 excites
  7. L18:1609 excites
  8. L9:750 excites
  9. L13:2758 excites
  10. L12:2564 excites
  11. L10:908 excites
  12. L9:2093 excites
  13. L14:2092 excites
  14. L17:2861 excites
  15. L10:2133 excites
  16. L15:1697 excites
  17. L9:619 excites
  18. L11:1973 excites
  19. L11:925 excites
  20. L15:931 excites
  21. L16:1983 excites
  22. L10:1430 weakly excites
  23. L11:2786 weakly excites
  24. L14:275 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.

2 values between -1.39 and -1.191 values between -1.00 and -0.808 values between -0.80 and -0.6117 values between -0.61 and -0.41178 values between -0.41 and -0.221292 values between -0.22 and -0.031982 values between -0.03 and 0.17338 values between 0.17 and 0.3658 values between 0.36 and 0.5616 values between 0.56 and 0.756 values between 0.75 and 0.953 values between 0.95 and 1.143 values between 1.34 and 1.532 values between 1.53 and 1.721 values between 2.11 and 2.311 values between 6.19 and 6.39offnegativepositiveoften

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. relations
  2. juni
  3. relationships
  4. relationships
  5. recursively
  6. relaciones
  7. 僭
  8. ance
  9. {}.
  10. angu

Suppresses

  1. erais
  2. .eth
  3. çon
  4. ắn
  5. -pres
  6. deen
  7. ерь
  8. 不见
  9. 粜
  10. -sided