Neuron Explorer

Qwen3-0.6B

'and' in lists of three

A neuron in layer 7 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 "and" when it comes just before the last item in a list, such as "assets, liabilities, and equity." It is negative on an opening bracket after a name, as in "Tiger (Panthera Tigris)." Turning it up did not make the writing clearly about lists.

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

    . It’s a common phrase used for a myriad of symptoms affecting your ability to think. Brain fog can include things like memory problems, lack of mental clarity, and an inability to focus, and put thoughts into words. Some people describe it as

  2. #2 peak

    from the seafood business. Industry personnel joined conservationists, researchers, and government officials in discussing a range of sustainability issues - marine spatial planning, traceability shortcomings, and next-generation aquaculture among them.⏎The toughest item on the agenda may well

  3. #3 peak

    able is essential. This comprehensive guide aims to provide a clear understanding of what a bilan comptable entails, including its key components such as assets, liabilities, and equity. By delving into the importance of the bilan comptable for financial

  4. #4 peak

    what they read with the help of these easytouse reading comprehension cards! Your students will build critical reading skills by selecting the meanings of metaphors, similes and more.⏎- Perfect for center activities, homework and filling downtime between lessons⏎-

  5. #5 peak

    , men were believed to grow beards for different reasons than men do today. The three main reasons prehistoric men had beards were for warmth, intimidation, and to defend their face from hits. In Egyptian times, men would dye their be

  6. #6 peak

    , messages, and information like recipes or instructions. In many cases, content like photos or videos is shared. Read on

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 can't. It's not a good idea. I have to think about it. So, in the end, I see a lot of people who are doing the same thing, and I can't do it. So, I have

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

    ye National Park.⏎Royal Manas National Park comprises three regions:⏎Fifty-eight mammal species have been reported in the park to date such as:⏎- Tiger (Panthera Tigris),⏎- Clouded Leopard (Neofelis

  2. #2 peak

    no sex, or no television. Afterwards, the students answered questions about their sexual attitudes and beliefs.⏎Taylor used video clips from six television programs: "Friends" (NBC), "Just Shoot Me" (NBC), "Ally McBeal

  3. #3 peak

    south. It is connected to the Atlantic Ocean via the narrow Strait of Gibraltar to the west.⏎In order, the five largest islands in the Mediterranean are Sicily (Italy), Sardinia (Italy), Cyprus, Corsica (France) and

  4. #4 peak

    Gratis is used as both an adjective (The drinks were gratis) and an adverb (Drinks were served gratis).

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. L6:597 suppresses
  2. L5:947 weakly suppresses
  3. L3:2501 weakly excites
  4. L5:509 weakly excites the word 'even' after a clause
  5. L6:1064 weakly excites
  6. L6:171 weakly excites
  7. L3:1959 weakly suppresses
  8. L5:756 weakly excites
  9. L5:891 weakly suppresses
  10. L3:522 weakly suppresses
  11. L5:744 weakly suppresses
  12. L3:1266 weakly suppresses
  13. L2:1169 weakly excites
  14. L4:1931 weakly excites
  15. L0:316 weakly suppresses
  16. L5:1356 weakly excites
  17. L5:345 weakly excites
  18. L6:2608 weakly suppresses
  19. L4:27 weakly excites
  20. L2:1124 weakly suppresses
  21. L5:2194 weakly suppresses
  22. L5:197 weakly excites
  23. L4:854 weakly excites
  24. L6:824 weakly excites

Feeds later layers

  1. L10:782 suppresses
  2. L8:1047 excites
  3. L8:986 suppresses
  4. L9:532 suppresses
  5. L13:912 suppresses
  6. L11:130 suppresses
  7. L12:1509 suppresses
  8. L12:690 suppresses
  9. L15:270 suppresses
  10. L14:1190 suppresses
  11. L16:1307 suppresses
  12. L11:76 excites
  13. L8:72 excites
  14. L13:106 excites
  15. L16:1141 suppresses
  16. L12:2811 excites
  17. L16:2421 suppresses
  18. L12:276 excites
  19. L15:1095 suppresses
  20. L12:683 excites
  21. L11:244 excites
  22. L14:1116 excites 'and' inside lists
  23. L11:644 excites
  24. L9:694 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 -2.86 and -2.751 values between -2.64 and -2.532 values between -2.43 and -2.321 values between -2.21 and -2.101 values between -2.10 and -1.991 values between -1.99 and -1.891 values between -1.78 and -1.671 values between -1.67 and -1.562 values between -1.56 and -1.461 values between -1.46 and -1.351 values between -1.35 and -1.242 values between -1.13 and -1.024 values between -1.02 and -0.922 values between -0.92 and -0.815 values between -0.81 and -0.7010 values between -0.70 and -0.5915 values between -0.59 and -0.4930 values between -0.49 and -0.3852 values between -0.38 and -0.27167 values between -0.27 and -0.16570 values between -0.16 and -0.061636 values between -0.06 and 0.05888 values between 0.05 and 0.16356 values between 0.16 and 0.27116 values between 0.27 and 0.3831 values between 0.38 and 0.483 values between 0.48 and 0.591 values between 0.59 and 0.702 values between 0.70 and 0.811 values between 1.13 and 1.242 values between 1.24 and 1.352 values between 1.35 and 1.45offnegativepositiveoften

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. unately
  2. 莎
  3. cams
  4. outfile
  5. шли
  6. 况
  7. }.
  8. 这时候
  9. contre
  10. 甸

Suppresses

  1. 人格
  2. ழ
  3. แรง
  4. 든
  5. iage
  6. revisit
  7. ide
  8. 行为
  9. icted
  10. иона