Neuron Explorer

Qwen3-0.6B

period in a middle initial

A neuron in layer 7 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 full stop after a person's middle initial, like the dot in 'William R. Pogue'. The activations are fairly weak, and turning it up gave text that does not obviously relate to names.

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

    table, ' get its DataSet. Dim thisDataSet As DataSet = CType(DataGrid1.DataSource, DataSet) ' Get the DataTableCollection. Dim tablesCollection As DataTableCollection = thisDataSet.Tables ' Get a specific table by name. Dim table As DataTable

  2. #2 peak

    The Mercury 7 Astronauts⏎William R. Pogue orbited the Earth for 84 days during the final manned flight

  3. #3 peak

    Dakota Game, Fish and Parks Department Authors are James R. Johnson and Gary E. Larson⏎$19.95 and Free Shipping

  4. #4 peak

    The President of the United States has seven roles in his job, from being the Commander-In-Chief to the Guardian of the Economy.⏎Brief biography of Gerald R. Ford, the 38th president of the United States of America.⏎Timeline

  5. #5 peak

    While taking some time in the middle to form the Industrial Workers of the World. Debs died in 1926.⏎5: Sources " Eugene V. Debs | Biography

  6. #6 peak

    to Thomas Aquinas, to Mortimer J. Adler. Now Adler instructs the area within the "uncommon good judgment"

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 "s" and "r" in the game of "Sudoku". So, I'm going to take the lead with a new game. I have a new puzzle board, and I want to play. Now, I'm going

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

    Mar. 1, 2007).⏎In 2006, the Department of Homeland Security (DHS)’s Customs and Border Protection apprehended 101,952 unaccompanied children (ali

  2. #2 peak

    can now be a threat to world peace. Finally, the U.S. government decision to adopt unilateral measures, bypassing the United Nations and the International Court of Justice, has jeopardized the slow, tentative process of creating a

  3. #3 peak

    promoting the rule of law, ensuring access to justice, and strengthening institutions. It is the intersection of SDG 11 and SDG 16 that holds importance, and hope, for vulnerable populations in fragile communities around the world

  4. #4 peak

    waters. Last year both governments signed a revised Great Lakes Water Quality Agreement and once again asked the IJC to report regularly on the health of the

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. L6:16 weakly excites
  2. L6:13 weakly excites
  3. L4:1308 weakly excites
  4. L3:2492 weakly suppresses
  5. L6:2834 weakly excites
  6. L2:923 weakly excites second half of 'side by side' and 'face-to-face'
  7. L3:1561 weakly excites
  8. L6:1970 weakly suppresses
  9. L1:1687 weakly excites
  10. L6:8 weakly suppresses
  11. L2:1490 weakly excites
  12. L5:1194 weakly suppresses
  13. L6:1520 weakly suppresses
  14. L3:1431 weakly excites
  15. L3:512 weakly suppresses
  16. L1:1441 weakly suppresses
  17. L3:2732 weakly excites
  18. L6:12 weakly excites
  19. L1:1343 weakly suppresses
  20. L4:99 weakly excites
  21. L5:742 weakly excites 'this' after a preposition
  22. L0:1808 weakly excites
  23. L4:21 weakly excites
  24. L2:2710 weakly excites

Feeds later layers

  1. L8:446 excites
  2. L10:1241 excites
  3. L12:2009 excites
  4. L20:900 excites end-of-text right after cut-off text
  5. L8:224 suppresses
  6. L11:1181 excites
  7. L9:853 weakly suppresses
  8. L8:121 weakly excites
  9. L8:2187 weakly suppresses
  10. L12:2131 weakly suppresses
  11. L19:619 weakly excites
  12. L12:246 weakly excites
  13. L18:151 weakly excites
  14. L18:195 weakly excites
  15. L8:1917 weakly excites
  16. L8:1998 weakly excites
  17. L16:2339 weakly suppresses
  18. L19:2146 weakly suppresses
  19. L16:2724 weakly excites
  20. L21:1337 weakly suppresses
  21. L8:530 weakly suppresses
  22. L13:450 weakly excites
  23. L14:1847 weakly excites
  24. L18:2909 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 -1.15 and -1.111 values between -0.72 and -0.681 values between -0.63 and -0.582 values between -0.58 and -0.534 values between -0.53 and -0.492 values between -0.49 and -0.4411 values between -0.44 and -0.3924 values between -0.39 and -0.3449 values between -0.34 and -0.2986 values between -0.29 and -0.25136 values between -0.25 and -0.20221 values between -0.20 and -0.15323 values between -0.15 and -0.10440 values between -0.10 and -0.06492 values between -0.06 and -0.01552 values between -0.01 and 0.04461 values between 0.04 and 0.09372 values between 0.09 and 0.14268 values between 0.14 and 0.18183 values between 0.18 and 0.23130 values between 0.23 and 0.2863 values between 0.28 and 0.3340 values between 0.33 and 0.3826 values between 0.38 and 0.426 values between 0.42 and 0.478 values between 0.47 and 0.524 values between 0.52 and 0.571 values between 0.57 and 0.611 values between 0.71 and 0.76offnegativepositiveoften

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. 因为你
  2. ue
  3. eval
  4. lest
  5. uen
  6. kj
  7. acionales
  8. 竞
  9. ::$
  10. ecz

Suppresses

  1. -dot
  2. 딩
  3. arding
  4. par
  5. átis
  6. 墉
  7. ERGY
  8. ptime
  9. romium
  10. .Has