Neuron Explorer

Qwen3-0.6B

word before him or he

A neuron in layer 26 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 a word that is usually followed by a pronoun such as "him", like "led", "tells", "earned" or "enables" in "led him away". It is slightly negative on opening quote marks. Turning it down or up changed the text in unrelated ways, so steering does not confirm the pattern.

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

    ing Spiritualism and Materialism’, Ludwig Andreas Feuerbach wrote- “Der Mensch ist, was er ißt.” (Man is what he eats) and by this,

  2. #2 peak

    can choose his own life. Man has dignity; he is intrinsically and individually valuable. He owns himself and the products of his labor. His rational nature enables him to build a free society where he is not a slave to his rulers.

  3. #3 peak

    most senior position.⏎Fisher wrote to Miller on 14 May 1912 offering him the position of Governor. Miller accepted and his Commission appointing him to the role was signed on 29 May 1912

  4. #4 peak

    in Venice, where his father was in practice. He was destined for the law, but his inherited tastes led him away from jurisprudence to medicine. He practised in several different towns

  5. #5 peak

    .⏎When Hamlet is engaged in a discussion with Gertrude, they are sparring with words. She is talking about one thing, he another. She tells him he has offended his father very much (meaning Claudius, who is not

  6. #6 peak

    feelings with his northern neighbor. He nevertheless won a major victory over the Edomites. Afterward, however, his veneration of the gods of Edom earned him a prophetic condemnation.⏎Seeking to punish Israel for raids against Judean

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: It sounds illogical that you could get you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you

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

    social scientist, and cognitive psychologist whose primary research interest was decision-making within organizations and is best known for the theory of “bounded rationality” and the theory of “satisficing.” He applied psychology to statistics, operations, and decision-making scenarios

  2. #2 peak

    a violation of the most trivial rules.4 Another characteristic of this crushing form of economic Darwinism is that it thrives on a kind of social amnesia that erases critical thought, historical analysis, and any understanding of broader systemic relations. In

  3. #3 peak

    the Oregon National Primate Research Center at Oregon Health & Science University, monitored the reproductive systems of healthy female nonhuman primates following exposure to Delta-9-tetrahydrocannabinol, or THC, the main psychoactive ingredient in marijuana

  4. #4 peak

    Genetic algorithms used to be one of the cornerstones of what is called computational intelligence. Computational intelligence is a subfield of artificial intelligence based mainly on the idea of creating intelligence by mimicking natural processes. For example, swarms display a certain kind

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. L24:477 strongly excites
  2. L24:2385 strongly excites
  3. L23:656 strongly suppresses 'the' in historical names
  4. L23:1052 suppresses
  5. L23:646 excites
  6. L24:440 suppresses
  7. L25:310 excites
  8. L23:3026 suppresses
  9. L21:1552 suppresses
  10. L23:346 suppresses
  11. L16:1238 suppresses
  12. L23:956 excites
  13. L22:520 excites
  14. L20:2754 weakly suppresses
  15. L25:2922 weakly suppresses
  16. L25:1971 weakly suppresses
  17. L22:2253 weakly excites
  18. L16:31 weakly suppresses
  19. L20:1446 weakly excites
  20. L16:39 weakly suppresses
  21. L16:1347 weakly suppresses
  22. L24:942 weakly suppresses
  23. L16:2464 weakly excites
  24. L23:2536 weakly suppresses

Feeds later layers

  1. L27:578 strongly suppresses
  2. L27:1244 strongly suppresses
  3. L27:737 weakly excites
  4. L27:717 weakly suppresses
  5. L27:486 weakly excites
  6. L27:1068 weakly excites
  7. L27:2412 weakly suppresses
  8. L27:1155 weakly suppresses just before a pronoun for God
  9. L27:544 weakly excites
  10. L27:1099 weakly excites
  11. L27:2190 weakly suppresses
  12. L27:1276 weakly excites
  13. L27:2641 weakly excites
  14. L27:2472 weakly excites
  15. L27:106 weakly excites
  16. L27:2972 weakly excites
  17. L27:3015 weakly excites 'F' starting a word in a split
  18. L27:2101 weakly excites
  19. L27:1399 weakly excites
  20. L27:2701 weakly excites
  21. L27:2135 weakly excites
  22. L27:1504 weakly suppresses
  23. L27:990 weakly suppresses
  24. L27:760 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.

4 values between -8.16 and -6.981 values between -6.98 and -5.807 values between -5.80 and -4.6315 values between -4.63 and -3.4528 values between -3.45 and -2.27122 values between -2.27 and -1.101440 values between -1.10 and 0.082203 values between 0.08 and 1.2655 values between 1.26 and 2.438 values between 2.43 and 3.6110 values between 3.61 and 4.791 values between 4.79 and 5.963 values between 5.96 and 7.142 values between 7.14 and 8.321 values between 8.32 and 9.493 values between 9.49 and 10.671 values between 10.67 and 11.851 values between 11.85 and 13.021 values between 13.02 and 14.201 values between 36.55 and 37.731 values between 37.73 and 38.91offnegativepositiveoften

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. orate
  2. viders
  3. rowsing
  4. 姐妹
  5. IFT
  6. orsk
  7. ее
  8. 坞
  9. eway
  10. 妇科

Suppresses

  1. he
  2. him
  3. 他
  4. 他在
  5. He
  6. 他也
  7. 他对
  8. 他说
  9. 他又
  10. he