Neuron Explorer

Qwen3-0.6B

'to' after 'in order for' phrases

A neuron in layer 3 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 comes after a named subject, as in 'in order for a region to be approved' or 'expect there to be'. Turning it up led to rambling, repetitive text rather than anything clearly connected to this 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

    .2: Variable Relationships⏎- For each pair of variables, do you expect there to be a relationship? That is, do you think a change in one variable is

  2. #2 peak

    it was not nearly as successful as it could have been. Or as successful as we, the outsiders looking in, would like for it to have been.⏎The Inventor's Mindset⏎One thing that should be stated

  3. #3 peak

    IR 1458+10B, in a paper in the current issue of The Astrophysical Journal.⏎"At such temperatures we expect the brown dwarf to have properties that are different from previously known brown dwarfs and much closer to those

  4. #4 peak

    icultural Area”. AVA’s are official wine grape growing regions of the United States designated by the Alcohol and Tobacco Tax and Trade Bureau. In order for a region to be approved as an AVA, it must be able to prove its identity as

  5. #5 peak

    a reference to the fact that they can form when no storms are present, and even though they usually occur under a cloud bank, it is possible for the weather to be very calm and even fairly sunny. It is rare for such a watersp

  6. #6 peak

    radiation. In order for the layers to adhere to each other, the light must pass through the layer being cured and into

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 first line, the second line, the third line, the fourth line. So many things to do. So many things to do. So many things to do. So many things to do. So many things to do. So many

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

    . To the unprofessional eye the object might look good, but the truth lies in the details: use of the correct reversible materials; intrusive interventional methods kept to a minimum; careful documentation of the research and the work process before, during and

  2. #2 peak

    is a year round affair. Earlier this year we committed ourselves to becoming an anti racist borough and this work supports how we are working hard to keep to that promise.⏎We want to present this important part of British history as an essential

  3. #3 peak

    risk of breast cancer and prostate cancer and also of the spread of the cancer to other places. This is according to a very recent paper (3) from Tulane University in New Orleans. Studies of cultured cancer cells and human cancers grafted to

  4. #4 peak

    changed their minds about her. They considered her knowledgable and wise and they seeked her council. The Owl no longer gave her advice, she kept it to herself and was sad for the folly of her kind.⏎View cards

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. L1:906 weakly suppresses
  2. L1:1506 weakly excites
  3. L0:2141 weakly excites
  4. L2:2949 weakly excites
  5. L0:613 weakly excites
  6. L2:1189 weakly suppresses
  7. L1:1380 weakly suppresses
  8. L1:2175 weakly suppresses
  9. L2:1899 weakly excites
  10. L2:3017 weakly excites
  11. L1:874 weakly suppresses
  12. L0:1068 weakly excites
  13. L2:1208 weakly suppresses
  14. L0:1551 weakly suppresses
  15. L1:124 weakly suppresses
  16. L0:686 weakly suppresses sentence-starting 'This' at text start
  17. L1:753 weakly excites
  18. L1:2682 weakly suppresses
  19. L2:111 weakly suppresses
  20. L0:678 weakly suppresses
  21. L1:988 weakly suppresses
  22. L1:2816 weakly suppresses
  23. L1:2792 weakly suppresses
  24. L0:1431 weakly excites

Feeds later layers

  1. L4:374 excites
  2. L14:2204 excites
  3. L5:53 excites
  4. L4:1089 weakly suppresses
  5. L18:2794 weakly suppresses
  6. L15:1412 weakly excites
  7. L15:2137 weakly excites
  8. L7:1217 weakly excites
  9. L8:256 weakly excites
  10. L4:212 weakly excites
  11. L10:372 weakly excites
  12. L11:723 weakly suppresses
  13. L14:198 weakly excites
  14. L15:2227 weakly suppresses
  15. L11:1311 weakly suppresses
  16. L14:822 weakly excites
  17. L14:1740 weakly excites
  18. L13:2015 weakly excites
  19. L13:1715 weakly suppresses
  20. L20:1074 weakly suppresses
  21. L12:60 weakly excites
  22. L11:2467 weakly excites
  23. L24:2637 weakly suppresses
  24. L16:1832 weakly 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 -2.21 and -2.011 values between -1.59 and -1.381 values between -1.38 and -1.172 values between -0.96 and -0.763 values between -0.76 and -0.5526 values between -0.55 and -0.34462 values between -0.34 and -0.132239 values between -0.13 and 0.081049 values between 0.08 and 0.2993 values between 0.29 and 0.495 values between 0.49 and 0.705 values between 0.70 and 0.917 values between 0.91 and 1.122 values between 1.12 and 1.332 values between 1.33 and 1.531 values between 1.53 and 1.742 values between 1.74 and 1.951 values between 1.95 and 2.162 values between 2.37 and 2.581 values between 2.99 and 3.201 values between 4.66 and 4.861 values between 5.28 and 5.491 values between 5.91 and 6.11offnegativepositiveoften

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. happen
  2. 挛
  3. bü
  4. 构
  5. 来讲
  6. oci
  7. 生物学
  8. unj
  9. \Migrations
  10. ander

Suppresses

  1. eful
  2. thalm
  3. edith
  4. orrow
  5. łość
  6. ivor
  7. 쓰
  8. newList
  9. rollment
  10. idth