Neuron Explorer

Qwen3-0.6B

'to' after verbs like continue or help

A neuron in layer 8 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 'to' when it follows verbs such as 'continue', 'help', or 'set', as in 'continue to reverberate'. It goes the other way on 'been' after 'has' or 'had'. Turning it up gave unrelated text, so the steering results do not show the pattern.

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

    ⏎(Nanowerk News) Around half a million people in the EU need treatment for serious skeletal defects every year. New research on bone tissue and repair promises to overcome inherent problems when a large bone graft is required.⏎Therapy of large

  2. #2 peak

    ricated. Regular maintenance helps detect wear and tear early, preventing small issues from escalating into major repairs. This proactive approach to maintenance extends the life of the HVAC

  3. #3 peak

    United States. The meeting took place on July 19-20, 1848, at Seneca Falls, New York. It helped to launch the U.S. campaign for women’s suffrage, or the right to

  4. #4 peak

    <|endoftext|>These days computer-generated fractal patterns are everywhere, from squiggly designs on computer art posters to illustrations in the most serious of physics journals. Interest continues to

  5. #5 peak

    natural resources consciously and sustainably.⏎The shocking images and news of the disastrous flooding in the Eifel, North Rhine-Westphalia and Bavaria continue to reverberate. They have once again raised our awareness of the fact that climate

  6. #6 peak

    : We have already seen a global reduction in air pollution due to lockdown and less travel. If we still use our cars less after the pandemic, this could continue to make a huge difference.⏎But, how much of a difference are we talking?⏎

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 thinking about lately about the value of my life, and I am going to make a decision. So far, this is the first time I have been thinking about it. I have been thinking about it. So I am going

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

    January or February with one-third of the tree's total annual nitrogen requirements. The amount depends on the tree's size and how many years it has been in the ground. A mature tree that has been planted 6 or more years

  2. #2 peak

    decide what other people can do with it. Copyright laws make it easier for authors to make money by selling their works. Because of copyright, a work can only be copied if the owner of the copyright gives permission.⏎When someone copies or edits a

  3. #3 peak

    happy or sad in the garden? Why?⏎- Who had the giant been visiting?⏎- How long had the giant been away?⏎- In your own words describe what the giant did when he arrived at

  4. #4 peak

    sensitive,’ ‘too emotional,’ and that their behaviors are either ‘too dramatic’ or ‘too timid.’⏎People who feel more deeply and intensely than others are more aware of subtleties; their brain processes information and reflects on it more deeply

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. L7:681 excites 'from' after prevent or stop
  2. L6:914 excites
  3. L6:439 suppresses
  4. L6:2747 suppresses
  5. L6:2359 excites
  6. L6:1133 excites "to" before an infinitive verb
  7. L4:518 suppresses
  8. L6:1621 suppresses
  9. L4:113 suppresses
  10. L6:2252 excites
  11. L6:1269 excites
  12. L7:6 weakly suppresses
  13. L4:521 weakly suppresses
  14. L6:1880 weakly excites "by" introducing a deadline or means
  15. L6:1103 weakly suppresses
  16. L6:2297 weakly excites
  17. L4:1394 weakly suppresses
  18. L2:1979 weakly excites
  19. L6:95 weakly suppresses
  20. L6:2091 weakly suppresses
  21. L6:1751 weakly excites
  22. L4:2216 weakly excites
  23. L3:755 weakly excites
  24. L5:1505 weakly suppresses

Feeds later layers

  1. L9:459 suppresses
  2. L12:2176 suppresses 'of' after a place or role noun
  3. L20:113 excites
  4. L9:222 suppresses
  5. L9:188 suppresses 'of' in 'National Institute of' names
  6. L19:1582 suppresses "of" after an abstract noun
  7. L13:1100 suppresses
  8. L18:1645 weakly excites
  9. L10:459 weakly suppresses
  10. L21:215 weakly excites
  11. L17:1225 weakly suppresses
  12. L9:52 weakly excites
  13. L24:85 weakly suppresses
  14. L17:2057 weakly suppresses
  15. L16:987 weakly suppresses
  16. L9:335 weakly suppresses 'of' after nouns like anniversary or onset
  17. L15:317 weakly excites
  18. L9:841 weakly suppresses
  19. L15:696 weakly suppresses
  20. L14:672 weakly suppresses
  21. L14:2275 weakly suppresses
  22. L18:1253 weakly suppresses
  23. L20:480 weakly suppresses
  24. L13:700 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 -3.24 and -3.101 values between -3.10 and -2.971 values between -2.84 and -2.711 values between -2.44 and -2.313 values between -2.31 and -2.171 values between -2.17 and -2.042 values between -2.04 and -1.911 values between -1.91 and -1.779 values between -1.77 and -1.647 values between -1.64 and -1.515 values between -1.51 and -1.3810 values between -1.38 and -1.2414 values between -1.24 and -1.1111 values between -1.11 and -0.9828 values between -0.98 and -0.8433 values between -0.84 and -0.7143 values between -0.71 and -0.5839 values between -0.58 and -0.4579 values between -0.45 and -0.31212 values between -0.31 and -0.18615 values between -0.18 and -0.051880 values between -0.05 and 0.09618 values between 0.09 and 0.22177 values between 0.22 and 0.3556 values between 0.35 and 0.4924 values between 0.49 and 0.6213 values between 0.62 and 0.756 values between 0.75 and 0.883 values between 0.88 and 1.023 values between 1.02 and 1.153 values between 1.15 and 1.283 values between 1.28 and 1.421 values between 1.42 and 1.552 values between 1.55 and 1.682 values between 1.81 and 1.951 values between 1.95 and 2.08offnegativepositiveoften

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. �能
  2. ensor
  3. 不容
  4. abric
  5. emain
  6. 不留
  7. ource
  8. roller
  9. ьер
  10. Uri

Suppresses

  1. 打扫
  2. gang
  3. ],'
  4. movement
  5. /documents
  6. (TAG
  7. /fwlink
  8. ],"
  9. (!(
  10. %; ⏎