Neuron Explorer

Qwen3-0.6B

sentence-ending full stops in web text

A neuron in layer 9 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 full stops that end a sentence in web pages and articles, especially after a name or a title. Its negative side barely matters. Turning it up or down changed the topic of the output, but not in a clear way.

Written by Claude from this neuron's activations and steering outputs. Confidence: low. Think it is wrong? Flag it (opens an email to the maker with the neuron already filled in).

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

    -blame/failure/low self-esteem Sadness/depression: loss . In Building Self-Esteem in Children Who Are Adopted or Fostered, Dr. Sue offers simple and practical advice to those supporting children aged 7+ to help

  2. #2 peak

    yellow wire is burned and the wire on the right side is deep inside of the cavity of the oven, the spade prong was stuck on the old element. The connection at the right pull off with a pair of pliers and just needs

  3. #3 peak

    devices and you want to represent them in JAVA. Current states of a class’s co… In this chapter, we will look into the concepts - Classes and Objects. Java class objects exhibit the properties and behaviors defined by its class. Java Objects.

  4. #4 peak

    As an application developer, you are free to use any of these ports. If definitions are helpful to you, use these vocabulary terms in order to get you started. Administer digital networking port b10 worksheet b17. Apple macos

  5. #5 peak

    when you register 4.⏎Discussion questions for ‘The Death and Life of the Great Lakes’⏎Skip to search form Skip to main content You are currently offline. Some features of the site may not work correctly. They are home to 3

  6. #6 peak

    by zero. Even on our own wiki, at one point.⏎Handwriting is such serious business that some people actually consider it an art to make words on paper. Never mind what words they actually wrote, whether they used printing or cursive is

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 number of the phone. A number is a 10-digit number, which is a 10-digit number. It is a 10-digit number, which is a 10-digit number, which is a 1

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

    ⏎The main crack<|endoftext|>analysis of algoritms⏎robert.kern at gmail.com⏎Fri Sep 10 00:10:22 CEST 2010⏎On 9/9/10

  2. #2 peak

    17:55 EST⏎- ASU instrument on NASA rover helps identify outcrop of long-sought

  3. #3 peak

    that has been in use worldwide for over 200 years. It has been available on the<|endoftext|>Register New Player⏎Welcome to our world of fun trivia quizzes and quiz games:⏎The Battle of Badr⏎"The Battle of Badr was

  4. #4 peak

    initial batch of 41 country pages, with more to follow. They are:⏎Afghanistan, Angola, Argentina, Bahrain, Belarus, Belgium, Burkina Faso, Burundi, Cambodia, Cameroon, Colombia, Costa Rica, Czech

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. L8:802 excites
  2. L8:625 weakly suppresses
  3. L7:1053 weakly suppresses
  4. L5:102 weakly excites
  5. L5:2422 weakly suppresses
  6. L8:324 weakly suppresses
  7. L7:58 weakly suppresses
  8. L7:1418 weakly excites
  9. L7:133 weakly suppresses line break inside a sentence
  10. L2:772 weakly excites
  11. L7:1040 weakly suppresses
  12. L2:1873 weakly excites
  13. L1:998 weakly excites
  14. L7:2976 weakly suppresses
  15. L3:2483 weakly excites
  16. L3:2973 weakly suppresses
  17. L5:2332 weakly suppresses
  18. L1:2877 weakly suppresses
  19. L3:2312 weakly excites
  20. L8:52 weakly excites
  21. L5:223 weakly excites
  22. L5:1026 weakly suppresses
  23. L4:2735 weakly suppresses
  24. L3:1790 weakly excites

Feeds later layers

  1. L18:1105 excites
  2. L15:1820 excites
  3. L14:2497 excites
  4. L13:1086 excites
  5. L15:2188 excites
  6. L15:1619 excites
  7. L10:258 excites
  8. L10:348 excites
  9. L15:1448 excites
  10. L15:1884 excites
  11. L12:195 excites
  12. L16:2634 weakly excites
  13. L10:97 weakly excites
  14. L15:2887 weakly excites
  15. L20:2082 weakly suppresses
  16. L21:530 weakly suppresses
  17. L12:1144 weakly excites
  18. L23:2491 weakly suppresses
  19. L13:437 weakly excites end of a short definition or claim
  20. L14:501 weakly excites
  21. L12:1725 weakly suppresses
  22. L14:285 weakly excites
  23. L14:548 weakly excites
  24. L12:959 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.

7 values between -0.79 and -0.7025 values between -0.70 and -0.6254 values between -0.62 and -0.5490 values between -0.54 and -0.45178 values between -0.45 and -0.37239 values between -0.37 and -0.28425 values between -0.28 and -0.20634 values between -0.20 and -0.12786 values between -0.12 and -0.03769 values between -0.03 and 0.05363 values between 0.05 and 0.14194 values between 0.14 and 0.2276 values between 0.22 and 0.3135 values between 0.31 and 0.3910 values between 0.39 and 0.474 values between 0.47 and 0.562 values between 0.56 and 0.646 values between 0.64 and 0.731 values between 0.73 and 0.812 values between 0.81 and 0.891 values between 0.89 and 0.981 values between 0.98 and 1.062 values between 1.06 and 1.151 values between 1.23 and 1.311 values between 1.40 and 1.481 values between 1.57 and 1.651 values between 2.49 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. мед
  2. enn
  3. enne
  4. ญา
  5. 甑
  6. Convertible
  7. -aged
  8. Perl
  9. |
  10. Verb

Suppresses

  1. $⏎
  2. arten
  3. $ ⏎
  4. $⏎
  5. $")⏎
  6. swana
  7. bjerg
  8. @⏎
  9. uces
  10. __()⏎