Neuron Explorer

Qwen3-0.6B

'that' after 'added' or 'adding'

A neuron in layer 6 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 lights up on the word 'that' when it comes right after someone 'added' or 'adding' something, as in 'He added that...'. When it was turned up, the written text did not clearly change in a way we could link to this habit.

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

    constant disappointment."⏎He added that technology now thrusts those alternate plans into our faces, which can make some people

  2. #2 peak

    shouldn't be eating or drinking.⏎Bener agreed it's difficult to know if consumption of sugary drinks has been falling much, adding that schools are only half the battle.⏎"Getting them<|endoftext|>The island of Guam is

  3. #3 peak

    access to information and communication technologies (ICTs) can mitigate many issues in SIDS and help to diversify their economies.⏎She added that innovations such as artificial intelligence, blockchain, drones, and mobile money were enabling progress

  4. #4 peak

    as it is today,” Bonde said, adding, "There were probably still people in the area and was probably a little bit marshy.”⏎He also added that the country's laws permit property owners to keep fossils discovered on their premises.⏎Talking

  5. #5 peak

    rely on focus groups of children to see what they like and what interests them. They have kid-testers participating in test sessions throughout the process. But she added that the team also spends a lot of time researching the developmental needs, social needs,

  6. #6 peak

    the butterflies feed or lay eggs,” said Rodrigues, adding that the Southern Birdwing (Troides minos), the biggest known butterfly in

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 a 3000000000 (A) style work of art. In this case, the title of the sculpture, the name of the work of art, is the same as the artist's self

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

    said. “Still, you realize all rocket launches have some inherent risk, so, it’s just hold your breath and hope everything goes well.”Go Back to Top. Skip To: Start of Article.<|endoftext|>David Woods of Ohio State and Robin

  2. #2 peak

    poverty is represented mainly by achievements in education and health. The Human Development Index (HDI) that is regularly estimated by the United Nations Development Programme (UN

  3. #3 peak

    home for about 15 years. She struggles with some behavioral problems at the home and has difficulty communicating her needs to the staff.⏎Dr. Smith remarks, “I notice that Stephanie has never had a pap smear before, at least according to

  4. #4 peak

    in the past few years, said Deborah McCullough, a professor of entomology and forestry at Michigan State University. Based on her research, she says: “There is no reason for a landscape ash tree to die from emerald ash

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. L5:1949 weakly excites
  2. L5:1003 weakly excites
  3. L3:1528 weakly excites
  4. L4:1484 weakly excites
  5. L2:2581 weakly excites
  6. L4:1349 weakly suppresses
  7. L1:1074 weakly excites
  8. L3:1679 weakly excites
  9. L5:881 weakly suppresses
  10. L3:1135 weakly excites
  11. L4:2058 weakly excites
  12. L1:722 weakly excites
  13. L2:1344 weakly suppresses
  14. L2:2064 weakly excites
  15. L4:2399 weakly suppresses
  16. L4:2898 weakly suppresses
  17. L2:410 weakly suppresses
  18. L3:2090 weakly suppresses
  19. L4:1104 weakly suppresses
  20. L5:316 weakly excites
  21. L2:253 weakly excites
  22. L4:1928 weakly excites unclear: personal names, mostly Indian
  23. L3:473 weakly excites
  24. L1:2895 weakly suppresses

Feeds later layers

  1. L9:1558 excites
  2. L7:983 excites
  3. L8:1929 excites
  4. L7:1028 excites
  5. L9:1900 excites
  6. L8:200 excites
  7. L9:2247 excites
  8. L9:2922 excites
  9. L8:2271 excites
  10. L10:653 suppresses
  11. L11:1876 excites
  12. L10:809 weakly excites
  13. L10:1517 weakly excites
  14. L14:2553 weakly suppresses
  15. L16:1757 weakly excites
  16. L7:1611 weakly excites
  17. L13:1569 weakly excites
  18. L8:1877 weakly suppresses
  19. L7:2815 weakly excites
  20. L21:3069 weakly suppresses
  21. L19:1807 weakly suppresses
  22. L11:2403 weakly excites
  23. L20:1666 weakly suppresses
  24. L18:464 weakly suppresses

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 -1.36 and -1.212 values between -1.21 and -1.052 values between -1.05 and -0.902 values between -0.90 and -0.755 values between -0.75 and -0.6012 values between -0.60 and -0.44117 values between -0.44 and -0.29643 values between -0.29 and -0.141713 values between -0.14 and 0.011081 values between 0.01 and 0.17243 values between 0.17 and 0.3236 values between 0.32 and 0.4714 values between 0.47 and 0.6210 values between 0.62 and 0.774 values between 0.93 and 1.084 values between 1.08 and 1.233 values between 1.23 and 1.381 values between 1.38 and 1.531 values between 1.53 and 1.692 values between 1.69 and 1.843 values between 1.84 and 1.992 values between 2.14 and 2.302 values between 2.30 and 2.451 values between 2.60 and 2.751 values between 3.51 and 3.671 values between 3.82 and 3.971 values between 4.27 and 4.431 values between 4.58 and 4.73offnegativepositiveoften

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. zac
  2. ovic
  3. 亲切
  4. 原谅
  5. DidLoad
  6. 话语
  7. 饼干
  8. breadcrumb
  9. requested
  10. olic

Suppresses

  1. SES
  2. comes
  3. SE
  4. endeavor
  5. TURE
  6. standing
  7. ?f
  8. YPE
  9. hir
  10. atem