Neuron Explorer

Qwen3-0.6B

'followed' and 'searched' after a subject

A neuron in layer 18 of Qwen3-0.6B. Its strongest snippets are mixed: they peak on different tokens. It is featured, so it has steering outputs.

What this neuron does

This neuron fires on verbs like 'followed', 'following', 'searched' and 'dreaming' that come right after a subject, as in 'scientists had searched'. Its strongest negative cases are truncation markers like '[...]' or '…' that cut a sentence short. Turning it up made the model drift into abstract, vague statements.

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

    . Only 183 seconds long, the simple melody and poetry captured the wounded hopefulness of its moment - and transcended its time to inspire generations that followed.⏎Charles J. Shields traces the song's origins - from the fire-bomb

  2. #2 peak

    Do not look directly at oncoming headlights, rather look toward the right side of the road and use the white fog line as a guide. When headlights from vehicles following you reflect in your rearview mirror, use the “day-night” feature on

  3. #3 peak

    ela, a professor at the University of California, Santa Barbara and the former spokesperson of the CMS experiment at the Large Hadron Collider.⏎For decades, scientists had searched for the elusive Higgs boson: the holy grail of modern physics and

  4. #4 peak

    was begun on a holiday at Braemar in Scotland at the prompting of Fanny’s young son. Kidnapped and The Master of Ballantrae followed on as his international fame as<|endoftext|>“Paradise Found and Lost” from

  5. #5 peak

    .⏎You may also like..⏎- Education and Training⏎- Special Populations⏎- Youth in Transition<|endoftext|>White Turkey – White Turkey⏎No, you’re not dreaming – White turkey are a real thing! White turkeys are bred from stock that

  6. #6 peak

    politics. The Ohio-born Hayes followed a fairly typical trajectory to the White House, checking off many of the most common

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: how to do a small ...??????????????????????????????????????????

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

    aphylaxis-like symptoms previously should always have injectable epinephrine (an Epi-Pen) nearby.⏎This answer should not be considered medical advice...This answer should not be considered medical advice and should not take the place of a

  2. #2 peak

    rather than cell references) from the ANOVA output as well as the Excel T... function. (Do not use the output from the ANOVA Template.xlsx file

  3. #3 peak

    ayanova, the Murray B. Sachs Professor in the Department of Biomedical Engineering at The Johns Hopkins University Schools of Engineering and Medicine and the project’s principal investigator… Continue reading.⏎Natalia Trayanova, a professor

  4. #4 peak

    bites are not initially painful and can go unnoticed for hours or days. This allows bed bugs to withdraw human blood for up to 10 minutes with each feeding...One bed bug will usually take more than one bite. Once a bed bug inserts

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. L15:1978 strongly suppresses
  2. L3:990 excites
  3. L16:64 excites
  4. L17:360 suppresses
  5. L16:60 suppresses
  6. L0:2889 suppresses
  7. L16:572 excites
  8. L16:493 weakly suppresses
  9. L16:155 weakly excites
  10. L16:110 weakly suppresses
  11. L0:2471 weakly suppresses
  12. L0:1964 weakly excites enzyme and protein biochemistry terms
  13. L16:215 weakly suppresses
  14. L10:273 weakly suppresses
  15. L6:944 weakly suppresses
  16. L16:83 weakly suppresses
  17. L17:765 weakly suppresses
  18. L16:162 weakly excites
  19. L0:2872 weakly suppresses
  20. L2:2028 weakly excites
  21. L16:75 weakly suppresses
  22. L16:260 weakly excites
  23. L1:24 weakly excites
  24. L2:2029 weakly suppresses

Feeds later layers

  1. L23:641 suppresses
  2. L20:653 excites the 9 in years like 19xx
  3. L27:325 suppresses
  4. L19:1004 excites
  5. L19:902 excites
  6. L26:362 weakly excites
  7. L20:2507 weakly excites
  8. L26:626 weakly suppresses
  9. L21:334 weakly excites
  10. L26:889 weakly excites
  11. L27:2834 weakly excites
  12. L22:384 weakly excites
  13. L26:791 weakly excites
  14. L26:2878 weakly suppresses
  15. L23:3011 weakly excites
  16. L19:572 weakly excites
  17. L20:1228 weakly excites
  18. L19:557 weakly excites
  19. L20:2475 weakly excites
  20. L27:2912 weakly excites
  21. L22:1084 weakly excites
  22. L23:270 weakly suppresses
  23. L26:2249 weakly excites
  24. L20:2782 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.

1 values between -42.94 and -41.771 values between -32.43 and -31.261 values between -26.59 and -25.422 values between -6.73 and -5.562 values between -5.56 and -4.392 values between -4.39 and -3.224 values between -3.22 and -2.0643 values between -2.06 and -0.892420 values between -0.89 and 0.281348 values between 0.28 and 1.4582 values between 1.45 and 2.622 values between 2.62 and 3.79offnegativepositiveoften

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. pu
  2. glitter
  3. bie
  4. قو
  5. grinder
  6. 赎
  7. Toe
  8. ilig
  9. atism
  10. lass

Suppresses

  1. 続きを読む
  2. edException
  3. ynchron
  4. clamation
  5. ../
  6. 征
  7. artial
  8. enko
  9. 心意
  10. ple