Neuron Explorer

Qwen3-0.6B

open parenthesis in citations

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 fires on an opening bracket "(" in reference lists and citations, like the bracket after an author's name and before a year. Turning it up did not produce a clear theme in the replies.

Written by Claude from this neuron's activations and steering outputs. Confidence: medium. Think it is wrong? Flag it (opens an email to the maker with the neuron already filled in).

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

    .⏎McLaren, P., & Farahmandpur, R. (2001). Class, Cultism, and Multiculturalism. Mult

  2. #2 peak

    provides an example of a revised “show-me” sentence and six exercises for students to practice.|⏎Baines, L., & Kunkel, A. (2010). Going bohemian

  3. #3 peak

    "Related research" (further below) or search for a different version of it.⏎Volume (Year): 42 (2011)⏎Issue (Month): 2 (June)⏎|Contact details of provider:|| Web page

  4. #4 peak

    set list for Large Scale Biosphere-Atmosphere .⏎Sign In above to download companion files.⏎Sign In above to download data set files.⏎|Data File (Granule)||File Size||Dates||

  5. #5 peak

    photography studio that Dalmau opened in 1854.⏎Download data is not yet available.⏎How to Cite⏎Cuenca, C. (2019). The The “Animated Polystereoscope

  6. #6 peak

    ||Classical and Quantum Gravity|⏎|State||Published - Jul 7 2003|⏎ASJC Scopus subject areas⏎- Physics and Astronomy (miscellaneous)<|endoftext|>Most computer users are familiar with the concept of spyware —

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: if I had a problem with a 100% rate of error in the assumption that the number of solutions of the quadratic equation in a variable with a constant term of $ \frac{2}{\cos \theta} $,

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

    of capital, or they will invest elsewhere.⏎The following is the WACC calculation formula:⏎WACC = E/V × Re + D/V × Rd × (1 - Tc)⏎Re = cost of equity⏎Rd = cost of debt

  2. #2 peak

    behavior of a population is known at all. * How to Use Slovin's Formula⏎Slovin's formula is written as:⏎n = N / (1 + Ne^2)⏎n = Number of samples⏎N = Total population⏎

  3. #3 peak

    m (18,510 ft) on its western summit and 5,6

  4. #4 peak

    ) you will obtain your answer:⏎F'(x) = 2x√(1+x6)⏎Comment. There is no need to integrate and then differentiate by x

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. L4:64 excites
  2. L3:57 excites
  3. L5:64 weakly excites
  4. L5:275 weakly excites
  5. L0:386 weakly excites
  6. L0:1166 weakly excites
  7. L1:2088 weakly suppresses
  8. L4:2230 weakly excites
  9. L3:2241 weakly excites
  10. L5:913 weakly suppresses
  11. L1:2444 weakly suppresses
  12. L4:469 weakly excites
  13. L5:739 weakly suppresses
  14. L2:2376 weakly excites
  15. L4:2797 weakly suppresses
  16. L3:2655 weakly suppresses
  17. L4:2866 weakly suppresses the word 'average' before a quantity
  18. L1:1464 weakly excites
  19. L5:2248 weakly suppresses
  20. L4:1957 weakly suppresses
  21. L1:922 weakly suppresses
  22. L4:794 weakly excites
  23. L3:207 weakly suppresses
  24. L0:1271 weakly excites

Feeds later layers

  1. L16:995 weakly suppresses
  2. L7:1981 weakly suppresses
  3. L22:2720 weakly excites
  4. L9:1496 weakly excites
  5. L8:273 weakly suppresses
  6. L17:562 weakly suppresses
  7. L11:606 weakly excites
  8. L13:1871 weakly excites
  9. L13:1145 weakly suppresses
  10. L10:646 weakly excites
  11. L13:236 weakly suppresses
  12. L17:545 weakly excites
  13. L11:552 weakly excites
  14. L21:1029 weakly excites
  15. L7:759 weakly excites
  16. L20:1090 weakly suppresses
  17. L19:1915 weakly excites
  18. L13:2825 weakly suppresses
  19. L13:1633 weakly excites
  20. L15:2858 weakly excites
  21. L17:1495 weakly excites 'without' introducing a contrast
  22. L9:1754 weakly excites
  23. L7:1374 weakly excites
  24. L7:2193 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 -0.60 and -0.552 values between -0.55 and -0.495 values between -0.49 and -0.4323 values between -0.43 and -0.3742 values between -0.37 and -0.3287 values between -0.32 and -0.26138 values between -0.26 and -0.20294 values between -0.20 and -0.14486 values between -0.14 and -0.09702 values between -0.09 and -0.03835 values between -0.03 and 0.03581 values between 0.03 and 0.09334 values between 0.09 and 0.15188 values between 0.15 and 0.2091 values between 0.20 and 0.2648 values between 0.26 and 0.3225 values between 0.32 and 0.3810 values between 0.38 and 0.445 values between 0.44 and 0.491 values between 0.55 and 0.611 values between 0.61 and 0.671 values between 0.78 and 0.841 values between 0.84 and 0.902 values between 0.90 and 0.952 values between 0.95 and 1.011 values between 1.42 and 1.471 values between 1.65 and 1.71offnegativepositiveoften

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. less
  2. inar
  3. INAL
  4. inos
  5. tb
  6. bb
  7. heets
  8. cest
  9. aul
  10. ports

Suppresses

  1. owego
  2. chiếu
  3. engage
  4. amous
  5. eri
  6. 可靠性
  7. vem
  8. rogue
  9. 疲劳
  10. 使用寿命