Neuron Explorer

Qwen3-0.6B

'in' right after a place or time

A neuron in layer 25 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 'in' when it comes straight after a place name or a time phrase, like 'located in', 'Jallianwala Bagh in' or 'April in'. Turning it up or down changed the model's text in ways that did not clearly relate to this word.

Written by Claude from this neuron's activations and steering outputs. Confidence: low. 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

    world as it was then and is even now.<|endoftext|>A NASA satellite the size of a small piano shot into space Wednesday atop an air-launched rocket, kicking off a two

  2. #2 peak

    plagiarism?" And why is<|endoftext|>The last few months have brought national attention to Tulsa, Oklahoma, as plans were unveiled to begin digging there in April in suspected mass graves — locations where corpses may have been dumped almost a century ago,

  3. #3 peak

    to 200 members of the Royal Army of Sulu, who have holed up in the village of Lahad Datu for the past two weeks in order to press their historic claim to the land. The Philippine and Malaysian governments are

  4. #4 peak

    ukan⏎Lachit Was born in 1622 on the 24th of November. He was born in the Ahim empire, located in today

  5. #5 peak

    ius Cicero⏎IN JULY 51 b.c. the Roman statesman and orator Marcus Tullius Cicero arrived in Cilicia, in

  6. #6 peak

    , the so called Amritsar Massacre of 1919. The massacre took place in a public park called Jallianwala Bagh in the city of Amritsar where British Indian army’s Colonel Reginald D

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: have you had a chance to get to some of the more remote areas that have some natural gas reserves, like peñolesole, peñolesole, peñolesole, peñolesole, peñolesole, pe

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

    ó, another student of Bortnyik, or László Káldor. One of Nemes’ most famous design is the poster for the

  2. #2 peak

    age and other places, in Sicily and Corsica, in Spain at Tartessus, and even beyond the pillars of Hercules at Gadeira (Cádiz)".⏎In the time of King

  3. #3 peak

    inadequate access to water and sanitation facilities. Nigeria has pledged its commitment to meeting Sustainable Development Goal (SDG) 6, which is to “ensure access to water and sanitation for all” by 2030. But the National

  4. #4 peak

    ounder and then pick the black doll). Watching black children struggle with those questions is very difficult.⏎Showing the video in a racially mixed class is risky, and always provokes interesting and complicated responses from the students. Often the classes can be unpredictable,

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. L21:728 excites
  2. L19:2814 suppresses
  3. L20:2031 excites
  4. L7:2761 excites
  5. L19:2643 excites "is" in "what is" questions
  6. L21:825 suppresses
  7. L20:2460 excites
  8. L18:1586 excites
  9. L22:489 suppresses
  10. L23:594 excites
  11. L13:597 weakly excites
  12. L20:2335 weakly suppresses comma ending an 'if' clause
  13. L4:4 weakly suppresses
  14. L21:918 weakly excites
  15. L10:469 weakly suppresses 'of' after a noun phrase
  16. L20:219 weakly suppresses "the" after time or place prepositions
  17. L21:109 weakly excites
  18. L12:158 weakly excites
  19. L22:990 weakly suppresses
  20. L1:1262 weakly suppresses
  21. L1:1 weakly suppresses
  22. L15:2370 weakly excites
  23. L22:543 weakly suppresses
  24. L22:349 weakly suppresses

Feeds later layers

  1. L26:1104 strongly excites
  2. L27:656 strongly excites
  3. L26:1836 excites
  4. L26:1046 excites
  5. L27:1534 weakly excites
  6. L26:1804 weakly suppresses
  7. L27:887 weakly suppresses
  8. L27:891 weakly suppresses
  9. L26:2153 weakly suppresses
  10. L27:2175 weakly excites
  11. L26:787 weakly suppresses
  12. L26:613 weakly excites
  13. L27:2410 weakly suppresses
  14. L26:400 weakly excites
  15. L27:1314 weakly excites
  16. L26:937 weakly suppresses
  17. L26:3013 weakly suppresses
  18. L26:1128 weakly excites
  19. L26:1304 weakly suppresses
  20. L26:2755 weakly excites
  21. L26:2428 weakly suppresses
  22. L26:1673 weakly suppresses
  23. L26:1843 weakly suppresses
  24. L27:968 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 -13.89 and -12.941 values between -12.00 and -11.051 values between -11.05 and -10.101 values between -10.10 and -9.152 values between -8.21 and -7.261 values between -7.26 and -6.3111 values between -6.31 and -5.367 values between -5.36 and -4.4114 values between -4.41 and -3.4726 values between -3.47 and -2.5251 values between -2.52 and -1.57252 values between -1.57 and -0.622385 values between -0.62 and 0.32880 values between 0.32 and 1.2791 values between 1.27 and 2.2243 values between 2.22 and 3.1739 values between 3.17 and 4.1218 values between 4.12 and 5.0620 values between 5.06 and 6.0118 values between 6.01 and 6.969 values between 6.96 and 7.916 values between 7.91 and 8.855 values between 8.85 and 9.806 values between 9.80 and 10.757 values between 10.75 and 11.702 values between 11.70 and 12.641 values between 12.64 and 13.593 values between 13.59 and 14.541 values between 14.54 and 15.491 values between 15.49 and 16.431 values between 16.43 and 17.381 values between 18.33 and 19.282 values between 19.28 and 20.231 values between 23.07 and 24.02offnegativepositiveoften

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. what
  2. what
  3. What
  4. What
  5. whats
  6. WHAT
  7. .what
  8. 什么
  9. 是什么
  10. .What

Suppresses

  1. nerRadius
  2. viso
  3. ñ
  4. ativity
  5. regation
  6. ثن
  7. 口气
  8. groundColor
  9. HttpException
  10. 溟