Neuron Explorer

Qwen3-0.6B

opening parenthesis after a term

A neuron in layer 13 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

Fires on the opening bracket or colon that comes right after a term and is about to explain it, such as a word followed by its meaning in brackets. Turning it up gave repetitive or rambling text that does not connect clearly to this 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

    among their many strange⏎body-shifting traits. Leaders of Mexican fisheries and scientists⏎are banding together to figure out whether the diablos⏎rojos (red devils

  2. #2 peak

    decades to the Sun Belt, where homes are cheap and jobs are plentiful but climate risks mount by the day? Or, following the logic of Biden’s own order, should funds be directed to more resilient regions where people might move someday? If it

  3. #3 peak

    00⏎American History / Women's Studies/Gender Studies⏎View main book page⏎Table of Contents⏎Chapter 1. Sleep Is for Sissies: Elite Males as Paragons of Wakefulness⏎Chapter 2.

  4. #4 peak

    the pointers is larger, this value may be positive or negative.<|endoftext|>Think of your body like a car. Where the muscles are the shock systems and the skeleton (joints/bones) is the frame of the car. The job of the

  5. #5 peak

    is hoped can ultimately influence public health policy. Tibetan mothers increased the practice of breastfeeding during the project. However, traditional customs of supplementing breast milk with tsampa (barley porridge) are deeply engrained in the mothers and family advisors,

  6. #6 peak

    sessions (majalis) in meeting places known as zawiyas, khanqahs or tekke. They strive for ihsan (perfection of worship), as detailed in a hadith: “Ihsan

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 Aha! So I'm not sure, so I have to. Let me. So I have to. Let me. I'm. So. I'm. So. So. Let me. So. Let. So. So

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

    shown to be generally safe, citing statistics on how many Americans have died after the seasonal influenza vaccine in recent years.⏎Each year, more than 165 million Americans get the flu shot, and according to the U.S. vaccine adverse event

  2. #2 peak

    his scholarship was withdrawn for lack of funds. “I thought the world was coming to an end,” the 97-year-old says. Mary Downey ’41 agrees: “People lost everything back then, their jobs and their savings,

  3. #3 peak

    ysis of human red blood cells would occur if the cell were in .. solution A) isotonic B) hypertonic C) hypotonic D) none of these I think it's B. I'm not too sure though.⏎November 2

  4. #4 peak

    is not improved by medication or therapy.⏎Share this article⏎Depression is one of the most common forms of mental illness, affecting more than 350 million people worldwide, according to a Reuters video. Bob Holmes is one of them.

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. L12:2240 suppresses
  2. L11:2719 suppresses
  3. L7:2514 suppresses
  4. L6:1389 excites
  5. L10:271 excites
  6. L11:547 excites
  7. L11:2066 suppresses
  8. L12:2947 excites
  9. L9:1065 suppresses
  10. L8:933 suppresses words introducing an example or aside
  11. L9:1414 suppresses opening parenthesis after copyright or year
  12. L8:747 weakly excites
  13. L0:866 weakly suppresses
  14. L9:1370 weakly suppresses
  15. L7:2085 weakly suppresses
  16. L7:3058 weakly excites
  17. L6:990 weakly excites
  18. L6:428 weakly excites
  19. L7:508 weakly excites
  20. L12:2171 weakly suppresses
  21. L11:408 weakly excites
  22. L12:2293 weakly excites
  23. L11:2166 weakly excites
  24. L11:498 weakly excites

Feeds later layers

  1. L14:2755 suppresses
  2. L16:2663 suppresses
  3. L14:2820 suppresses "and" or "or" linking two counted items
  4. L15:2407 suppresses
  5. L16:3011 suppresses
  6. L17:1869 suppresses
  7. L15:2997 suppresses
  8. L16:2917 suppresses
  9. L14:2116 weakly suppresses
  10. L16:123 weakly excites
  11. L16:978 weakly suppresses
  12. L14:2006 weakly suppresses
  13. L15:2198 weakly suppresses
  14. L19:1125 weakly excites
  15. L16:2105 weakly suppresses
  16. L18:119 weakly excites
  17. L15:2720 weakly suppresses
  18. L22:1716 weakly suppresses
  19. L18:576 weakly suppresses
  20. L16:268 weakly excites
  21. L15:2499 weakly suppresses
  22. L15:1236 weakly suppresses
  23. L15:1547 weakly suppresses "also" before a verb
  24. L14:1933 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 -3.64 and -3.421 values between -2.98 and -2.761 values between -2.76 and -2.541 values between -2.54 and -2.323 values between -2.32 and -2.118 values between -2.11 and -1.8914 values between -1.89 and -1.6720 values between -1.67 and -1.4555 values between -1.45 and -1.2389 values between -1.23 and -1.01169 values between -1.01 and -0.80311 values between -0.80 and -0.58551 values between -0.58 and -0.36873 values between -0.36 and -0.141080 values between -0.14 and 0.08336 values between 0.08 and 0.30161 values between 0.30 and 0.5263 values between 0.52 and 0.7352 values between 0.73 and 0.9525 values between 0.95 and 1.1731 values between 1.17 and 1.3920 values between 1.39 and 1.619 values between 1.61 and 1.836 values between 1.83 and 2.0410 values between 2.04 and 2.263 values between 2.26 and 2.483 values between 2.48 and 2.701 values between 2.70 and 2.923 values between 2.92 and 3.141 values between 3.14 and 3.351 values between 3.35 and 3.571 values between 3.57 and 3.791 values between 3.79 and 4.011 values between 4.01 and 4.231 values between 4.45 and 4.671 values between 4.67 and 4.881 values between 4.88 and 5.10offnegativepositiveoften

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. agraph
  2. �
  3. 蛉
  4. -kind
  5. :⏎
  6. predicted
  7. ıyla
  8. bare
  9. extensively
  10. Scrollbar

Suppresses

  1. ains
  2. aining
  3. 制
  4. teness
  5. aways
  6. apsed
  7. AIN
  8. ระว
  9. /apt
  10. =j