Neuron Explorer

Qwen3-0.6B

capital L starting a place name

A neuron in layer 26 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

It fires on a capital L that starts a place or person's name, such as Lusaka, Lorraine, Lhasa or Lourdes. The opposite side is a lowercase l starting an ordinary word like 'lures'. Steering did not produce a clear theme.

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

    (Cape Town, South Africa) is a data analytics company using aerial imagery and machine learning algorithms to optimize crop performance for farmers globally.⏎AgriPredict (Lusaka, Zambia) provides farmers with access to information that will help them identify

  2. #2 peak

    one thousand dead and thirty to forty thousand jailed throughout Spain.⏎The day after the nationwide general strike was called on October 4, 1934, Lluís Companys, now president of the Generalitat, proclaimed the "

  3. #3 peak

    . In an earthquake, liquefaction can cause the ground to behave like quicksand, as seen in New Zealand and in the 1989 earthquake in Loma Prieta in the mountains of Santa Cruz, Calif.⏎In Japan, most

  4. #4 peak

    © CRDP de Lorraine, 2012).⏎The most outstanding examples of Renaissance architecture in

  5. #5 peak

    00,000 are ethnic Tibetans. In a survey carried out in Lhasa in July 1993 on the southern Lingkor, a

  6. #6 peak

    —on children is exceedingly rare in the Andes (except in cases of child sacrifice), so it is particularly revealing of the unstable sociopolitical conditions in the LIP [Late Intermediate Period] in the Ayacucho Basin.”⏎Previous work

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: there is not this rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule rule

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

    . They’re now one of the most common types of aquarium fish available, as well as being popular with anglers who enjoy catching them with live bait or artificial lures.⏎You might have heard of Synodontis catfish before, but chances

  2. #2 peak

    own species. Cally Harper, a doctoral student at Brown University, along with a team of researchers, set out to better understand how bats use their tongues when lapping up nectar. Scientists have known that they have hairy-looking tongues; the

  3. #3 peak

    9–1892) used both the one- and two-syllable pronunciations:⏎The purple flower droops; the golden bee⏎Is lily-cradled: I alone awake.⏎Both of these poets rhyme flower and

  4. #4 peak

    for honouring the country’s leaders. At many universities, North Korean students would wake their foreign classmates up and make them bring flowers to the bronze statue of Kim Il Sung in Wonsan to praise him. The foreign students were asked to give

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. L21:160 suppresses
  2. L22:245 suppresses
  3. L23:268 excites
  4. L23:789 excites
  5. L24:482 suppresses inside Italian names and places
  6. L18:247 suppresses
  7. L23:352 suppresses
  8. L20:2575 excites
  9. L24:624 suppresses
  10. L22:461 suppresses
  11. L23:1760 suppresses
  12. L21:252 excites
  13. L18:198 weakly suppresses
  14. L16:1 weakly suppresses
  15. L20:1186 weakly suppresses
  16. L22:2157 weakly excites
  17. L20:234 weakly excites
  18. L24:346 weakly excites
  19. L25:338 weakly suppresses
  20. L25:744 weakly excites
  21. L1:1981 weakly suppresses
  22. L5:315 weakly suppresses
  23. L3:1599 weakly suppresses
  24. L18:550 weakly suppresses

Feeds later layers

  1. L27:515 weakly excites
  2. L27:2610 weakly excites
  3. L27:1583 weakly suppresses
  4. L27:2731 weakly excites
  5. L27:2593 weakly excites unclear: first letter of a name fragment
  6. L27:181 weakly excites
  7. L27:1184 weakly excites
  8. L27:854 weakly excites
  9. L27:2503 weakly suppresses
  10. L27:822 weakly excites
  11. L27:547 weakly suppresses
  12. L27:2694 weakly excites
  13. L27:408 weakly excites
  14. L27:1250 weakly excites
  15. L27:1258 weakly excites
  16. L27:1415 weakly excites
  17. L27:2742 weakly suppresses 'w' inside split words
  18. L27:2730 weakly excites
  19. L27:2866 weakly excites
  20. L27:1661 weakly excites
  21. L27:580 weakly excites
  22. L27:2100 weakly excites
  23. L27:2492 weakly excites
  24. L27:1129 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 -19.56 and -18.711 values between -13.59 and -12.731 values between -11.03 and -10.172 values between -8.46 and -7.611 values between -5.90 and -5.051 values between -5.05 and -4.205 values between -4.20 and -3.344 values between -3.34 and -2.496 values between -2.49 and -1.6480 values between -1.64 and -0.782479 values between -0.78 and 0.071258 values between 0.07 and 0.9345 values between 0.93 and 1.786 values between 1.78 and 2.635 values between 2.63 and 3.493 values between 4.34 and 5.202 values between 5.20 and 6.053 values between 6.05 and 6.901 values between 7.76 and 8.611 values between 9.46 and 10.321 values between 12.03 and 12.882 values between 13.73 and 14.59offnegativepositiveoften

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. ozy
  2. meld
  3. ainter
  4. neys
  5. ocaust
  6. iteur
  7. ушки
  8. rence
  9. 身影
  10. 踏实

Suppresses

  1. orris
  2. samt
  3. занят
  4. يرا
  5. %E
  6. PROCUREMENT
  7. Specs
  8. -rule
  9. postage
  10. Beet