Neuron Explorer

Qwen3-0.6B

comma before a qualifying clause

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 commas in the middle of a sentence, right after a full statement, where the writer is about to add a qualifier. When it is pushed down, the model tends to write the word 'especially' over and over, so it seems to hold that word back.

Written by Claude from this neuron's activations and steering outputs. Confidence: high. 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

    are the serosa, muscularis externa, submucosa and mucosa. This layered arrangement follows the same general structure in all regions of the stomach, and throughout the entire gastrointestinal tract. The outer layer of the stomach wall is smooth

  2. #2 peak

    transfer protocol (HTTP) server that runs in the background as a daemon. Because⏎it is a daemon, it must do all the things that background applications do, namely,⏎read its configuration files, go into the background, shut down when told

  3. #3 peak

    with a spread of about 50 feet. The compound leaves give it a lacy, airy appearance. This tree will tolerate a wide range of soil conditions, but it does require good drainage. Leaves will turn orange to red, and sometimes

  4. #4 peak

    as long as it did or we would have suffered a gap.”⏎But Landsat 6 was developed and launched outside of NASA’s control by a provate company, and Irons

  5. #5 peak

    south pole, providing further evidence of precipitation.⏎Elizabeth Turtle of John Hopkins University in Laurel, Maryland, and colleagues compared images from two flybys of the moon, taken on 3 July 2004 and 6 June 2

  6. #6 peak

    useful, but what if you have to make several copies, or you just want to make a complete backup of a hard drive? Drive images let you put everything, and we mean everything, from your hard drive in one big file.⏎With an

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: especially, especially, especially, especially, especially, especially, especially, especially, especially, especially, especially, especially, especially, especially, especially, especially, especially, especially, especially, especially, especially, especially, especially, especially,

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

    a few of the many ways to use edtech:⏎1. Conquering Teacher Shortage and Limited Course Offerings.⏎Finding qualified teachers can be a challenge, especially at this point in the year, but that doesn’t have to stop you

  2. #2 peak

    further costs.⏎4. Challenges and Considerations:⏎a. Affordability: Medical insurance premiums can be costly, especially for those without employer-sponsored coverage. Government subsidies and tax credits can help make

  3. #3 peak

    inflamed, it can swell and fill with pus. This can cause severe pain, nausea, and vomiting. Swelling is therefore an important symptom to watch for, especially if it is accompanied by other symptoms of appendicit

  4. #4 peak

    ism. Some of the most famous ukiyo-e artists include Katsushika Hokusai,<|endoftext|>Weight loss can be a major struggle for some people, so much so that they turn to gastric bypass surgery to help them achieve their weight

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. L17:118 excites comma after a clause ending
  2. L14:35 suppresses
  3. L16:240 excites comma after 'sometimes' or 'the truth'
  4. L17:2004 excites
  5. L18:1523 excites
  6. L18:366 excites comma inside numbers and dates
  7. L18:2376 suppresses
  8. L18:569 excites
  9. L18:100 excites
  10. L9:77 excites commas in citations and name lists
  11. L11:154 suppresses commas after citation numbers
  12. L17:2836 weakly excites
  13. L3:71 weakly suppresses
  14. L10:612 weakly excites full stop ending boilerplate or a title
  15. L16:241 weakly excites
  16. L0:178 weakly suppresses
  17. L0:1224 weakly suppresses
  18. L16:1748 weakly suppresses
  19. L5:2309 weakly excites
  20. L16:431 weakly excites
  21. L0:1301 weakly suppresses
  22. L2:483 weakly suppresses
  23. L18:1869 weakly excites
  24. L14:2254 weakly suppresses

Feeds later layers

  1. L20:1519 strongly suppresses
  2. L26:2236 strongly suppresses comma before an 'especially' aside
  3. L27:2047 strongly suppresses
  4. L23:720 suppresses
  5. L22:368 suppresses
  6. L22:158 excites
  7. L24:203 excites
  8. L26:236 suppresses
  9. L27:358 suppresses
  10. L27:213 excites
  11. L20:2082 suppresses
  12. L26:113 excites
  13. L20:622 excites
  14. L22:940 suppresses
  15. L27:747 excites
  16. L27:86 excites
  17. L21:436 excites
  18. L24:216 suppresses
  19. L26:22 excites the 'res' fragment inside words
  20. L26:606 weakly excites
  21. L20:882 weakly excites
  22. L26:80 weakly excites
  23. L26:388 weakly excites
  24. L21:384 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.

2 values between -11.36 and -10.842 values between -9.80 and -9.283 values between -7.19 and -6.672 values between -6.67 and -6.152 values between -6.15 and -5.634 values between -5.11 and -4.591 values between -4.59 and -4.074 values between -4.07 and -3.554 values between -3.55 and -3.034 values between -3.03 and -2.5110 values between -2.51 and -1.9911 values between -1.99 and -1.4724 values between -1.47 and -0.9588 values between -0.95 and -0.431608 values between -0.43 and 0.101754 values between 0.10 and 0.62331 values between 0.62 and 1.1417 values between 1.14 and 1.664 values between 1.66 and 2.185 values between 2.18 and 2.708 values between 2.70 and 3.227 values between 3.22 and 3.743 values between 3.74 and 4.264 values between 4.26 and 4.781 values between 4.78 and 5.302 values between 5.82 and 6.351 values between 6.87 and 7.391 values between 7.39 and 7.911 values between 8.95 and 9.47offnegativepositiveoften

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. =
  2. {|
  3. iled
  4. =⏎⏎
  5. تها
  6. = ⏎
  7. {.
  8. =⏎
  9. |
  10. �

Suppresses

  1. especially
  2. 特别是
  3. 尤其是
  4. especially
  5. 尤其
  6. 尤其是在
  7. particularly
  8. 特别是在
  9. particularly
  10. Especially