Neuron Explorer

Qwen3-0.6B

"well" after "how"

A neuron in layer 6 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 the word "well" in phrases like "how well," asking about how effective or good something is. It also dips on the ends of some adverbs like "textually" or "sparsely." Turning it up did not produce any clear theme in the sample text.

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

    ority homes during the instructional period. We administered follow-up assessments in the fall after the instructional period ended and in the spring of the following year to determine how well students maintained and consolidated target academic words. Students in the intervention group maintained their relative

  2. #2 peak

    5 year old graduation<|endoftext|>Look at the egg⏎picture. Write the name of the egg part in the box. Click the 'check'⏎button to see how well you did. When finished⏎print the worksheet to learn more.⏎Thank you for

  3. #3 peak

    experiencing through the goggles. The researchers were testing a phenomenon called “path integration” to determine how well the body was able to keep track of itself and its movements within the physical plane

  4. #4 peak

    that this tailor-made Chinese intervention program is very effective for improving word reading fluency of Chinese ordinary students and those with reading difficulties. For factors that may affect how well children learn to read, it appears that school resources, student academic ability, and

  5. #5 peak

    up to a year on your maintenance dose to notice a true change. If the treatment is successful, you will continue maintenance shots for three to five years.⏎How well do these shots actually work?⏎As with any treatment, effectiveness will vary from person

  6. #6 peak

    cadmium in children’s products– today the agency backed down. They are suggesting that the industry self-enforce. Hmmmm? How well did that work out the last time when we were talking about lead. We are

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: I have been using the word "certain" in a way that's problematic for my social media. I have been using the word "certain" in a way that's problematic for my social media. I have been using the word

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

    This coursebook can be used to support any pedagogical or research activities in text technologies, the history of the book, the history of information, and textually based work in the digital humanities. Through careful explanations of the field, examinations of

  2. #2 peak

    most populous of India’s northeastern states, is an expansive green of rice paddies, bamboo, banana, and sal trees. Bamboo huts dot its sparsely populated hamlets, ringed by low hills in the distance. Along the narrow

  3. #3 peak

    and construction, complement this landscape. The intimate paths show a human scale appropriate for the private world the Abkhazis wanted to create for themselves.⏎The sculpturally strong Garry oaks predominate the site and provide a unifying sense

  4. #4 peak

    -bedded poorly sorted sand and gravel. The aquifer is composed of hydraulically connected alluvial and terrace deposits that unconformably overlie the Permian

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. L4:2743 suppresses
  2. L0:828 suppresses
  3. L4:521 suppresses
  4. L1:1049 excites
  5. L3:1774 suppresses
  6. L3:530 excites
  7. L4:1743 excites
  8. L2:2960 excites
  9. L5:125 suppresses "of" after all three, both, ten
  10. L2:2932 excites
  11. L2:166 suppresses
  12. L5:2841 excites
  13. L0:1383 excites
  14. L1:818 weakly excites
  15. L3:682 weakly excites
  16. L3:1717 weakly suppresses
  17. L1:1239 weakly suppresses
  18. L4:1763 weakly excites
  19. L4:2722 weakly excites
  20. L3:6 weakly suppresses
  21. L4:450 weakly suppresses
  22. L3:228 weakly suppresses
  23. L5:1214 weakly suppresses
  24. L5:2641 weakly excites unclear: scattered word fragments

Feeds later layers

  1. L8:252 strongly suppresses 'of' after 'each' or 'one' or 'all'
  2. L17:985 strongly excites
  3. L7:173 strongly excites
  4. L15:557 strongly suppresses
  5. L11:810 suppresses
  6. L8:2755 excites
  7. L10:167 excites
  8. L21:716 excites
  9. L18:602 excites
  10. L19:663 excites
  11. L18:217 suppresses
  12. L16:1433 excites
  13. L14:280 suppresses
  14. L12:1058 excites
  15. L20:519 suppresses
  16. L10:289 suppresses
  17. L21:274 suppresses
  18. L17:1694 suppresses
  19. L17:831 excites
  20. L22:404 excites
  21. L21:68 suppresses
  22. L7:2164 excites period inside country abbreviations
  23. L18:452 suppresses
  24. L17:199 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 -2.47 and -2.351 values between -1.87 and -1.751 values between -1.51 and -1.391 values between -1.39 and -1.271 values between -1.27 and -1.152 values between -1.03 and -0.911 values between -0.91 and -0.791 values between -0.67 and -0.552 values between -0.55 and -0.438 values between -0.43 and -0.31135 values between -0.31 and -0.19867 values between -0.19 and -0.071849 values between -0.07 and 0.05696 values between 0.05 and 0.17195 values between 0.17 and 0.2997 values between 0.29 and 0.4136 values between 0.41 and 0.5310 values between 0.53 and 0.652 values between 0.65 and 0.771 values between 0.77 and 0.891 values between 2.21 and 2.33offnegativepositiveoften

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. ouble
  2. 时候
  3. ones
  4. counterpart
  5. ほうが
  6. counterparts
  7. pieces
  8. 片子
  9. obstacle
  10. 時候

Suppresses

  1. 的话题
  2. ric
  3. olen
  4. ymm
  5. eut
  6. rics
  7. ERICAN
  8. CGRect
  9. entials
  10. 的人都