Neuron Explorer

Qwen3-0.6B

word before a number, range or quantity

A neuron in layer 22 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 words that are followed by a number or a measured amount, such as 'maintained at' before a pH value or 'the difference between' before a term. Turning it up made the model's writing drift into unrelated topics, with no obvious pattern.

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

    individual and physical examination must be performed in a respectful and culturally sensitive manner. Privacy is essential and the nurse must be aware of their own ...⏎You have reached your viewing limit for this book.⏎What people are saying - Write a review⏎Respir

  2. #2 peak

    your computer uses DNS to retrieve the website’s IP address of 2001:41c8:20::19.⏎We Focus On:⏎Behind

  3. #3 peak

    17. Draw a diagram showing optimum combination of consumer.⏎18. Mention any two characteristics of oligopoly market⏎19. Clarify the difference between average revenue and marginal revenue.⏎20. Calculate the ex-ante aggregate demand,

  4. #4 peak

    80 - 120 ppm. The pH is the measure of the acidity of the swimming pool water, and the pH level should be maintained at 7.2 - 7.6 ppm. The pH of pool water is

  5. #5 peak

    It can be watched in many languages and these can be access by clicking on the backwards and forwards buttons (that look like |< or >|) after you have pressed play.⏎Tiny Happy People is a website dedicated

  6. #6 peak

    as factors in Parkinson's pathophysiology, multiple articles describing nuances of deep brain stimulation, and stories about exceptional athletes living with Parkinson's.⏎Read on for all of November's Parkinson's news!⏎SCIENCE AND RESEARCH Updates⏎- Chu et

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: "rationala ravenasravena ravena ravena ravena ravena ravena ravena ravena ravena ravena ravena ravena ravena raven

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

    interdental cleaning such as flossing or using the Waterpik, can get lost in the hustle. The ugly truth is that this simple task gets put on many people’s low priority list and can be the very reason their oral health

  2. #2 peak

    of credibility. It would not be an exaggeration to say that from ancient times until now, there has not been a period in which this question has not been put to a writer or an artist without regard to his nationality or the motives for his

  3. #3 peak

    illustration of how one subject is put for another is found in comparing the two seemingly contradictory biblical accounts of the death of

  4. #4 peak

    all directions. This may help to make some furtive words more readable. Also the spacing, boldness and capital case can be switched to make the letters stand out from their context. Optionally a handy hover-cross is shown, highlighting those letters

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. L20:959 suppresses
  2. L21:237 excites
  3. L21:746 excites
  4. L18:569 suppresses
  5. L16:582 suppresses
  6. L19:2987 suppresses
  7. L17:152 suppresses mid-word fragments, often 'intric'
  8. L16:222 suppresses
  9. L5:50 suppresses
  10. L11:42 excites
  11. L21:475 suppresses
  12. L7:9 excites
  13. L7:110 excites
  14. L3:115 excites
  15. L2:75 excites
  16. L8:10 suppresses
  17. L18:396 excites
  18. L11:300 excites
  19. L10:1651 suppresses
  20. L13:1638 suppresses
  21. L6:13 suppresses
  22. L8:7 suppresses
  23. L18:871 suppresses
  24. L20:1285 excites

Feeds later layers

  1. L23:254 excites
  2. L27:160 suppresses
  3. L24:385 excites
  4. L25:206 excites
  5. L27:342 excites
  6. L26:495 excites
  7. L23:154 excites
  8. L26:685 suppresses
  9. L26:442 suppresses
  10. L24:34 suppresses
  11. L23:760 suppresses
  12. L23:630 excites
  13. L23:25 excites
  14. L25:1493 suppresses
  15. L24:2423 weakly suppresses
  16. L23:1613 weakly suppresses
  17. L24:105 weakly excites
  18. L27:50 weakly excites
  19. L26:1041 weakly suppresses
  20. L24:943 weakly suppresses
  21. L23:385 weakly suppresses
  22. L27:766 weakly suppresses word endings split mid-word, like 'shower' in 'showered'
  23. L27:792 weakly suppresses
  24. L26:210 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.

1 values between -22.11 and -21.302 values between -21.30 and -20.481 values between -20.48 and -19.671 values between -19.67 and -18.861 values between -18.86 and -18.042 values between -18.04 and -17.231 values between -17.23 and -16.413 values between -16.41 and -15.6010 values between -15.60 and -14.7918 values between -14.79 and -13.9723 values between -13.97 and -13.1626 values between -13.16 and -12.3533 values between -12.35 and -11.5355 values between -11.53 and -10.7275 values between -10.72 and -9.90113 values between -9.90 and -9.09143 values between -9.09 and -8.28162 values between -8.28 and -7.46209 values between -7.46 and -6.65207 values between -6.65 and -5.84242 values between -5.84 and -5.02291 values between -5.02 and -4.21315 values between -4.21 and -3.40292 values between -3.40 and -2.58310 values between -2.58 and -1.77269 values between -1.77 and -0.95221 values between -0.95 and -0.14232 values between -0.14 and 0.67181 values between 0.67 and 1.49148 values between 1.49 and 2.30117 values between 2.30 and 3.1161 values between 3.11 and 3.9349 values between 3.93 and 4.7431 values between 4.74 and 5.5624 values between 5.56 and 6.3712 values between 6.37 and 7.1814 values between 7.18 and 8.006 values between 8.00 and 8.813 values between 8.81 and 9.624 values between 9.62 and 10.44offnegativepositiveoften

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. Professional
  2. iterations
  3. setObject
  4. 梅州
  5. 轮融资
  6. Worksheets
  7. 发展格局
  8. Permissions
  9. 经开区
  10. Psychological

Suppresses

  1. nels
  2. zyć
  3. gado
  4. gio
  5. nement
  6. naire
  7. ties
  8. keiten
  9. thers
  10. rious