Neuron Explorer

Qwen3-0.6B

unclear: end markers and clause commas

A neuron in layer 15 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 is weak and mixed. It fires on document-end markers and on some words after an introductory phrase, and it is most negative on commas that close such a phrase. The steering outputs were repetitive and do not reveal a clear meaning.

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

    , all of which are dependent on biodiversity. They are predominantly public goods with limited or no markets and do not command any price in the conventional economic system, so their loss is

  2. #2 peak

    i>⏎;; Now insert one character at the beginning of the buffer: (goto-char (point-min)) ⇒ 1 (insert "Q") ⇒ nil⏎<|endoftext|>A bead is a small bauble or jewel with a hole through its middle for

  3. #3 peak

    After Effects using precomps!⏎Check back next week to learn the difference between frame rate and shutter speed, or view more lessons at the Vimeo Video School .<|endoftext|>Some time back I presented the untold story of Kebara 2, the

  4. #4 peak

    an error has occurred. The ldexpf() and ldexpl() functions do not set errno.⏎See attributes(5) for descriptions of the following attributes:<|endoftext|>The

  5. #5 peak

    tool. Follow this same procedure to use them. They take measurements fixed on their grid direction, even if the tool endpoints are not perfectly aligned to the grid.<|endoftext|>Khalistan Issue: A brief Analysis⏎Author: Isha Pandey⏎

  6. #6 peak

    and a highly centralized economic system caused Albania to lag far behind its neighbors in terms of economic development.⏎After Hoxha died in 1985, his hand-picked successor, Ramiz Alia, who became party leader while retaining

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 have been thinking about the ways in which the choice of a single person can be interpreted as a way of understanding the world in which we live. This is, in other words, a way of understanding the world as a single person,

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

    date now remembered as Kristal Nacht, or the Night of Broken Glass.⏎That night, as Nazis broke the windows of Jewish stores and burned their synagogues, Goldner’s uncle was arrested along with more than 20,00

  2. #2 peak

    ons of England organized numerous uprisings. In the face of such strong and well-organized opposition, King John was forced to renounce certain rights and to grant a charter of liberties

  3. #3 peak

    led armies at Budapest, despite nearly succeeding, they failed and thus the door was open to cross the Oder⏎River and assault Berlin. As this was happening, British and Canadian forces were slogging their way towards the Rhine (taking the

  4. #4 peak

    the governors around him. The kings of Babylon were exceedingly prideful and rash. They had little regard for life. In the midst of such evil kingdoms and governments, the gatekeepers of the Lord influenced kings, governors, and peoples. There is

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. L9:652 excites
  2. L14:1845 suppresses
  3. L14:1492 suppresses
  4. L12:2016 suppresses
  5. L13:1732 suppresses
  6. L12:1529 weakly excites
  7. L13:1226 weakly suppresses
  8. L14:1658 weakly excites
  9. L13:2051 weakly suppresses
  10. L13:1511 weakly excites
  11. L12:2994 weakly excites
  12. L12:2924 weakly suppresses
  13. L13:775 weakly excites
  14. L14:655 weakly excites
  15. L14:531 weakly suppresses
  16. L12:1536 weakly suppresses
  17. L11:2876 weakly suppresses
  18. L13:109 weakly suppresses
  19. L11:2879 weakly suppresses
  20. L14:2316 weakly excites
  21. L11:182 weakly excites
  22. L11:2517 weakly excites
  23. L8:1041 weakly suppresses
  24. L13:2465 weakly excites

Feeds later layers

  1. L16:2501 excites
  2. L16:2368 excites
  3. L16:939 excites
  4. L18:1724 excites
  5. L16:2988 weakly excites
  6. L22:2321 weakly excites
  7. L16:3043 weakly excites
  8. L18:1897 weakly excites
  9. L20:1205 weakly excites
  10. L16:2949 weakly excites
  11. L16:1674 weakly excites
  12. L18:1629 weakly excites
  13. L17:2322 weakly suppresses
  14. L17:3019 weakly excites
  15. L19:2858 weakly excites
  16. L18:2349 weakly excites
  17. L16:1527 weakly excites
  18. L16:890 weakly excites
  19. L16:2984 weakly excites
  20. L22:1477 weakly excites
  21. L23:1906 weakly excites
  22. L21:1645 weakly excites
  23. L20:1745 weakly excites
  24. L18:2736 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 -9.20 and -8.901 values between -4.20 and -3.904 values between -3.31 and -3.024 values between -3.02 and -2.7318 values between -2.73 and -2.439 values between -2.43 and -2.1427 values between -2.14 and -1.8437 values between -1.84 and -1.5553 values between -1.55 and -1.26106 values between -1.26 and -0.96151 values between -0.96 and -0.67303 values between -0.67 and -0.37708 values between -0.37 and -0.081775 values between -0.08 and 0.21433 values between 0.21 and 0.51178 values between 0.51 and 0.8074 values between 0.80 and 1.1014 values between 1.10 and 1.398 values between 1.39 and 1.682 values between 1.68 and 1.981 values between 1.98 and 2.271 values between 2.27 and 2.57offnegativepositiveoften

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. nhật
  2. rels
  3. 成
  4. иш
  5. [first
  6. elist
  7. aravel
  8. unft
  9. onna
  10. UICollectionViewCell

Suppresses

  1. eph
  2. eus
  3. strup
  4. 본
  5. 傧
  6. theValue
  7. 盲
  8. 了自己的
  9. .CreateDirectory
  10. alone