Neuron Explorer

Qwen3-0.6B

closing bracket after a list number or letter

A neuron in layer 19 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 a closing bracket that ends a list label, like the ")" in "A)", "1)" or "7)". It is turned down by closing brackets that end a word inside a sentence. Turning it up gave off-topic text, so the link to writing is unclear.

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

    roussonetia papyrifera_), of _hau_ (_Hilasens tiliasens_), etc.<|endoftext|>The American Meteorological Society⏎) promotes the development and dissemination of information and education on the atmospheric⏎Atmospheric

  2. #2 peak

    6 = 0 is a solution of the homogeneous equation A Z = 0. Answer the following in terms of V , W , and Z . a) Find some solution of A X = 3 Y 1 . Solution : X

  3. #3 peak

    What is the period of a point moving along a circular path of radius 5 cm at the linear speed of 200π cm / minutes?⏎A) 1<|endoftext|>Galactorrhea is a discharge of milk-like substance from the

  4. #4 peak

    ings.⏎The succulent thrives indoors in a well-drained container where bright sunlight bathes the plant.⏎Young plants should be watered Every day.⏎7) Ripsali bacciferra-⏎8) string of tears-⏎Pack of

  5. #5 peak

    particular, that the crystals were pure protein, and assembled by autocatalysis⏎) were incorrect. The first electron microscopical images of TMV were made

  6. #6 peak

    1; Basic Chemistrynbsp;nbsp;Chemistry Questions 8211; Learn Basic Chemistry Part 1 (1-25)nbsp;1) When a beam of white rays is dispersed by a prism which colour will be refr

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: I know that the case of the "chasing a dog" is a classic example of the case of the *non-mutation* of a right in the case of the case of the *non-macre* of a right. I

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

    relationship, describe it in non-statistical language, with a visual analysis of the means. Then, use Excel to create formulas for a Multiple Means (Fishers) comparison for the two means that differ the most. Show the Excel formulas that are

  2. #2 peak

    MONSEIGNEUR: I HAVE the honor to inform you that Florida and the Lucayan Islands are situate on one side of the Bahama (old) Channel, which passes between Havanna (Cuba) and Florida. But nearer

  3. #3 peak

    plexus|⏎|Muscularis externa||Smooth muscle layers (longitudinal, circular, oblique), myenteric (Auerbach’s) plexus|⏎|Serosa||Connective tissue, mesoderm

  4. #4 peak

    Peritoneal Cancer Cedars-Sinai⏎(2 days ago) Cancer that has spread to the lining surfaces of the peritoneal (abdominal) cavity from ovarian cancer, primary colorectal cancer, appendiceal cancer, or

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. L0:870 excites
  2. L2:1384 suppresses
  3. L4:516 suppresses
  4. L18:569 excites
  5. L15:324 excites
  6. L18:198 excites
  7. L16:97 excites
  8. L16:13 suppresses
  9. L16:145 excites
  10. L10:1052 suppresses
  11. L16:289 excites
  12. L0:304 suppresses
  13. L17:38 weakly suppresses
  14. L16:96 weakly excites
  15. L16:7 weakly excites
  16. L15:123 weakly suppresses
  17. L0:110 weakly excites
  18. L7:879 weakly excites
  19. L17:506 weakly suppresses
  20. L14:51 weakly excites colon that introduces a statement
  21. L8:14 weakly excites
  22. L18:110 weakly excites
  23. L8:10 weakly excites
  24. L16:15 weakly excites

Feeds later layers

  1. L26:381 excites
  2. L24:2198 excites closing parentheses in lists
  3. L26:362 excites
  4. L23:742 excites
  5. L27:1450 excites
  6. L20:112 suppresses
  7. L27:183 excites
  8. L22:1367 suppresses
  9. L26:193 weakly suppresses
  10. L20:558 weakly excites
  11. L21:576 weakly suppresses
  12. L26:2766 weakly suppresses
  13. L25:422 weakly suppresses
  14. L20:662 weakly suppresses
  15. L22:1084 weakly suppresses
  16. L21:2478 weakly excites
  17. L27:2206 weakly excites
  18. L20:1393 weakly suppresses
  19. L22:2435 weakly suppresses
  20. L26:1472 weakly suppresses
  21. L25:483 weakly excites
  22. L23:2070 weakly suppresses
  23. L25:273 weakly excites dollar sign before an amount
  24. L26:729 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 -9.71 and -9.291 values between -7.19 and -6.771 values between -6.35 and -5.931 values between -4.67 and -4.251 values between -4.25 and -3.831 values between -3.83 and -3.411 values between -3.41 and -2.991 values between -2.99 and -2.572 values between -2.57 and -2.152 values between -2.15 and -1.7317 values between -1.73 and -1.3198 values between -1.31 and -0.89336 values between -0.89 and -0.471321 values between -0.47 and -0.051997 values between -0.05 and 0.3797 values between 0.37 and 0.7914 values between 0.79 and 1.218 values between 1.21 and 1.632 values between 1.63 and 2.052 values between 2.05 and 2.471 values between 2.89 and 3.311 values between 4.57 and 4.992 values between 6.67 and 7.09offnegativepositiveoften

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. sonian
  2. etus
  3. ゞ
  4. dling
  5. rets
  6. ulario
  7. ruption
  8. pling
  9. extr
  10. isches

Suppresses

  1. иде
  2. vak
  3. trí
  4. Mach
  5. ystone
  6. atables
  7. getClass
  8. ạc
  9. Ủ
  10. bef