Neuron Explorer

Qwen3-0.6B

period after abbreviation in citations

A neuron in layer 5 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 reacts to the dot that follows a short abbreviation in a reference, like "Vol.", "Oct." or "var.", where a number or name is expected next. Turning it up made the model's text drift into unrelated topics, so the steering results do not show a clear theme.

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

    to all use restrictions contained in the Terms and Conditions of Use of the BioOne website.⏎Vol. 65 • No. 2<|endoftext|>FC: Eugene V. Debs

  2. #2 peak

    A central point; a point of concentration. 2⏎v. t. - To bring to a focus; to focalize; as, to focus a

  3. #3 peak

    london-history-1216214947983326-8 по слайдам⏎2008 by Mag. Andreas Hofer and the 4 DHWIM⏎It is thought that London

  4. #4 peak

    fractures, not less, in women because women have less bone<|endoftext|>WebMD Medical News⏎Daniel J. DeNoon⏎Louise Chang, MD⏎Oct. 9, 2012 – About 13,00

  5. #5 peak

    : Three varieties of this North American subspecies occur in BC. Fruits black or purplish-black; nutlets slightly wrinkled or pebbly var. melan

  6. #6 peak

    _html/wp-content/plugins/ultimate-google-analytics/ultimate_ga.php on line 524⏎GUJHS. 2007 March; Vol. 4, No. 1⏎Sophie Cowan, ‘0

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 problem is to find the value of this expression, and the answer is something. So, the problem is to find the value of this expression, and the answer is something. This is a problem, but it's not.

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

    40 percent of excess body weight. However, according to a new study, 21 percent of gastric bypass surgery patients develop a drinking problem.⏎Bariatric (weight loss) surgery has become popular among Americans as an easy weight loss solution

  2. #2 peak

    a point that Dr. Pretlow called “alarming.” But why? He noted the tendency, in some quarters, to jump to a solution like bariatric surgery, and rhapsodize about what an acceptable safe effective treatment alternative it

  3. #3 peak

    aparoscopic Gastric Bypass ( A combined restrictive and malabsorptive procedure)⏎With laparoscopic techniques, the bariatric surgeons staple across the top portion of the stomach to create a very small stomach pouch

  4. #4 peak

    patients were four times more likely to say they had received treatment for substance abuse when they hadn’t.⏎“This indicates that treatment programs are underutilized by bariatric surgery patients with alcohol

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. L2:1769 suppresses
  2. L3:699 suppresses
  3. L1:1728 excites
  4. L3:1161 suppresses first digit after 'more than'
  5. L1:1135 weakly suppresses
  6. L0:456 weakly suppresses
  7. L2:2883 weakly excites
  8. L3:51 weakly excites
  9. L3:1753 weakly excites
  10. L1:1439 weakly suppresses
  11. L0:1757 weakly excites
  12. L2:825 weakly excites
  13. L2:2495 weakly excites
  14. L0:847 weakly suppresses
  15. L0:2484 weakly excites
  16. L4:1013 weakly suppresses
  17. L4:1218 weakly suppresses
  18. L2:1005 weakly suppresses the '6' in years like 196x
  19. L4:869 weakly excites
  20. L1:2798 weakly excites
  21. L3:1915 weakly suppresses
  22. L2:868 weakly excites line breaks inside lists
  23. L4:2005 weakly suppresses
  24. L1:947 weakly suppresses

Feeds later layers

  1. L8:517 excites
  2. L13:2803 excites
  3. L7:765 excites
  4. L12:792 excites
  5. L9:897 weakly excites
  6. L15:998 weakly excites period inside an abbreviation
  7. L7:887 weakly suppresses
  8. L15:1976 weakly excites
  9. L9:2898 weakly excites
  10. L23:505 weakly suppresses
  11. L19:1337 weakly excites
  12. L16:2128 weakly excites
  13. L26:1888 weakly excites
  14. L6:82 weakly excites
  15. L7:346 weakly excites
  16. L6:1478 weakly excites
  17. L16:1319 weakly excites
  18. L14:2798 weakly excites
  19. L26:1391 weakly excites
  20. L10:62 weakly suppresses
  21. L7:2898 weakly excites
  22. L15:1117 weakly excites
  23. L13:2513 weakly excites
  24. L26:1599 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 -1.39 and -1.291 values between -1.29 and -1.201 values between -1.20 and -1.102 values between -1.00 and -0.913 values between -0.81 and -0.714 values between -0.71 and -0.6210 values between -0.62 and -0.5226 values between -0.52 and -0.4251 values between -0.42 and -0.33101 values between -0.33 and -0.23186 values between -0.23 and -0.13685 values between -0.13 and -0.041742 values between -0.04 and 0.06751 values between 0.06 and 0.16244 values between 0.16 and 0.2568 values between 0.25 and 0.3513 values between 0.35 and 0.445 values between 0.44 and 0.543 values between 0.54 and 0.641 values between 0.64 and 0.731 values between 0.73 and 0.831 values between 0.83 and 0.932 values between 0.93 and 1.021 values between 1.22 and 1.311 values between 1.60 and 1.701 values between 1.70 and 1.792 values between 2.37 and 2.47offnegativepositiveoften

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. 护肤品
  2. /legal
  3. intest
  4. /books
  5. 进出口
  6. 聯絡
  7. 肢
  8. .Buttons
  9. .’”⏎⏎
  10. ventus

Suppresses

  1. anvas
  2. plode
  3. rage
  4. झ
  5. زال
  6. -space
  7. ingly
  8. <algorithm
  9. oldown
  10. unce