Neuron Explorer

Qwen3-0.6B

'turn' in 'which in turn'

A neuron in layer 2 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', which links a cause to its knock-on effect. Turning it up did not make the model's text visibly use this phrase.

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

    . The trees that die as a result of rising water levels attract insects, which in turn feed woodpeckers, whose holes later provide homes for other wildlife.⏎Figure

  2. #2 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

  3. #3 peak

    film, including one in which the dwarfs build and carve<|endoftext|>Gothic architecture is a good architectural design which in turn was especially popular in Europe throughout the Middle Age range, lasting into the middle and

  4. #4 peak

    onions, have been shown to reduce inflammation in the body, which in turn lowers the risk of artery hardening. Fresh, peeled beets can be thinly

  5. #5 peak

    secure seal that prevents particle, vapor, and fluid leakages, which in turn protects and maintains pressure whilst keeping debris out, ensuring proper function and safety.<|endoftext|>

  6. #6 peak

    initial state of the atmosphere is critical. Even if Doppler radar information may seem highly localized, we find that it enhances the regional atmospheric conditions, which, in turn, can significantly improve the dynamic prediction of how the monsoon depression will move as

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 time that, with a few years' passage, to the equality of two variables... I think the argument of reverse the inequality, I am not so much the time as... and so. that the previous and subsequent, and equal,

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

    in one of its caves during the Arabic period. This tradition is reflected in the writings of Fray Antonio Agustín de Milla and Suazo, dated back to the 17th

  2. #2 peak

    , America embraces people of all faiths and of no faith. We are Christians and Jews, Muslims and Hindus, Buddhists and Sikhs, atheists and agnostics. Our religious diversity enriches our cultural fabric and reminds us that what binds

  3. #3 peak

    values - before there developed⏎the actual beginnings of the Philippine revolution that fell under the class leadership of the ilustrados or the liberal bourgeoisie⏎that surrounded Aguinaldo.⏎In this Propaganda Movement, Dr. Jose Rizal made

  4. #4 peak

    third of U.S. adults under age 30 are unaffiliated.⏎By the year 2050, the study projects, atheists, agnostics and religiously<|endoftext|>Sea Monster Eggs and Leadership⏎In the early 9

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. L0:2204 weakly excites
  2. L1:2968 weakly excites
  3. L1:14 weakly suppresses
  4. L0:2460 weakly suppresses
  5. L0:2712 weakly excites
  6. L0:49 weakly suppresses
  7. L1:6 weakly excites
  8. L1:164 weakly suppresses
  9. L0:1171 weakly excites
  10. L1:2253 weakly suppresses
  11. L0:2942 weakly suppresses
  12. L1:1699 weakly suppresses
  13. L1:2118 weakly excites
  14. L0:397 weakly suppresses
  15. L1:1866 weakly suppresses
  16. L1:2651 weakly excites
  17. L1:2357 weakly excites
  18. L0:1289 weakly suppresses
  19. L1:486 weakly excites
  20. L1:1583 weakly excites
  21. L1:1719 weakly excites
  22. L0:2841 weakly suppresses
  23. L1:2307 weakly suppresses
  24. L0:2099 weakly suppresses

Feeds later layers

  1. L10:2439 weakly excites
  2. L5:1789 weakly suppresses
  3. L4:1154 weakly excites
  4. L14:2050 weakly suppresses
  5. L16:2011 weakly excites
  6. L25:1922 weakly suppresses
  7. L18:899 weakly suppresses
  8. L4:646 weakly suppresses
  9. L27:2023 weakly suppresses
  10. L5:2330 weakly excites
  11. L21:369 weakly excites
  12. L16:606 weakly suppresses
  13. L16:2102 weakly suppresses
  14. L16:1644 weakly suppresses
  15. L17:1165 weakly excites
  16. L9:1237 weakly excites
  17. L22:36 weakly excites
  18. L7:2221 weakly suppresses
  19. L13:532 weakly suppresses
  20. L11:1910 weakly excites
  21. L11:1203 weakly excites
  22. L19:2410 weakly suppresses
  23. L13:2579 weakly suppresses
  24. L6:2423 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 -1.41 and -1.281 values between -1.28 and -1.143 values between -1.00 and -0.864 values between -0.72 and -0.583 values between -0.58 and -0.4414 values between -0.44 and -0.30191 values between -0.30 and -0.161140 values between -0.16 and -0.022024 values between -0.02 and 0.12458 values between 0.12 and 0.2547 values between 0.25 and 0.3910 values between 0.39 and 0.533 values between 0.53 and 0.671 values between 0.81 and 0.951 values between 0.95 and 1.091 values between 1.23 and 1.371 values between 3.31 and 3.451 values between 3.45 and 3.591 values between 3.59 and 3.731 values between 3.73 and 3.871 values between 3.87 and 4.011 values between 4.01 and 4.15offnegativepositiveoften

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. кат
  2. 覚
  3. nable
  4. bolt
  5. worthy
  6. .getHeader
  7. 檎
  8. ñe
  9. =((
  10. ient

Suppresses

  1. ursively
  2. DATES
  3. udios
  4. lexible
  5. 相见
  6. 羽毛
  7. psy
  8. етод
  9. inality
  10. arma