Neuron Explorer

Qwen3-0.6B

"to" or "for" linking into a following phrase

A neuron in layer 11 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 lights up on small linking words like "to" or "for" when they lead into the next part of a sentence, such as "to find", "to Mars" or "to himself". When it was turned up, the writing did not clearly change in any obvious way.

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

    others when using the ecosystem approach as a basis for projects and programmes. The searchable component of the ecosystem approach sourcebook⏎can be used in addition to this guide to find information on case studies and tools which have been used to meet the principles of

  2. #2 peak

    scientists suffer programmatic whiplash whenever the human spaceflight program is redirected; investigations of the Moon, say, may be abandoned in favor of robotic precursor missions to Mars

  3. #3 peak

    hissing and clawing), to, finally, that level of self-communion in which the husband will seek to explain his wife’s behavior and their crumbling marriage to himself.⏎Last, there is the narrator’s ambiguous silence, looking on dispass

  4. #4 peak

    chemical substance that contains aluminum sulfate and a second chemical, usually potassium sulfate. It is used to shrink tissues, to stop bleeding, and to boost the immune response to a vaccine.<|endoftext|>Wand Chhakna / Wand Ke Chhak

  5. #5 peak

    decide whether to<|endoftext|>Stunning new data not yet publicly released shows Louisiana losing its battle with rising seas much more quickly than even the most pessimistic studies have predicted to date.⏎While state officials continue to argue over restoration projects to save the state’s

  6. #6 peak

    you might return "(False, reason)". Often, you can use exceptions to handle this situation.⏎Sometimes you might want to return objects of different types based on arguments to the function. For instance, did the caller ask for the data in this format

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 know that the problem is to be solved with a right angle. The problem is to be solved with a right angle. The problem is to be solved with a right angle. The problem is to be solved with a right angle. The

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

    sold as fresh grapefruit.<|endoftext|>AC — see alternating current.⏎Alternating Current (AC) — A type of electrical current, the direction of which is reversed at regular intervals or cycles. Electricity transmission networks use AC because voltage can be controlled with

  2. #2 peak

    at:http://www.msfc.nasa.gov/news⏎The above story is based on materials provided by NASA's Marshall Space Flight Center. Note: Materials may be edited for content and length.⏎Cite This Page:<|endoftext|>By A. M. Sullivan

  3. #3 peak

    as play, as well as sleep and time to rest, are important factors influencing a child's growth and developmental outcomes.⏎Play is defined as engaging in an activity for enjoyment purposes and should come naturally to children.⏎But when a child is confined in

  4. #4 peak

    (in part or in whole) was last reviewed on 4 February 2012.⏎Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.3 or any

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. L9:161 excites
  2. L8:577 excites
  3. L9:1358 excites
  4. L6:1752 excites
  5. L10:1609 weakly suppresses
  6. L4:152 weakly excites
  7. L9:103 weakly excites "as" or "since" giving a reason
  8. L8:772 weakly suppresses "it" as a subject pronoun
  9. L7:1903 weakly excites
  10. L9:2726 weakly suppresses
  11. L9:1890 weakly excites
  12. L8:705 weakly excites
  13. L8:151 weakly excites
  14. L8:230 weakly excites
  15. L5:20 weakly suppresses
  16. L2:840 weakly suppresses
  17. L6:655 weakly excites
  18. L0:2685 weakly excites
  19. L7:337 weakly excites
  20. L4:801 weakly excites
  21. L8:115 weakly suppresses
  22. L10:132 weakly excites
  23. L6:447 weakly suppresses
  24. L2:2630 weakly suppresses

Feeds later layers

  1. L14:674 excites
  2. L13:918 excites
  3. L15:161 excites
  4. L16:749 excites
  5. L14:419 excites
  6. L18:1154 excites
  7. L17:964 excites
  8. L16:509 suppresses
  9. L16:1026 weakly excites
  10. L14:2349 weakly excites
  11. L14:1133 weakly excites unclear: commas and nouns before an aside
  12. L16:928 weakly excites
  13. L16:1315 weakly excites word or phrase being named or defined
  14. L15:1902 weakly suppresses
  15. L14:1025 weakly excites
  16. L15:2247 weakly excites
  17. L15:360 weakly excites
  18. L16:507 weakly excites
  19. L15:1041 weakly excites
  20. L13:961 weakly excites
  21. L14:660 weakly excites
  22. L19:1294 weakly excites
  23. L19:663 weakly excites
  24. L12:644 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 -2.27 and -2.141 values between -2.14 and -2.012 values between -1.74 and -1.612 values between -1.61 and -1.481 values between -1.48 and -1.354 values between -1.35 and -1.225 values between -1.22 and -1.097 values between -1.09 and -0.968 values between -0.96 and -0.8314 values between -0.83 and -0.6916 values between -0.69 and -0.5625 values between -0.56 and -0.4372 values between -0.43 and -0.30292 values between -0.30 and -0.17894 values between -0.17 and -0.041752 values between -0.04 and 0.09547 values between 0.09 and 0.23125 values between 0.23 and 0.3650 values between 0.36 and 0.4932 values between 0.49 and 0.6223 values between 0.62 and 0.759 values between 0.75 and 0.886 values between 0.88 and 1.014 values between 1.01 and 1.144 values between 1.14 and 1.282 values between 1.28 and 1.413 values between 1.41 and 1.542 values between 1.54 and 1.672 values between 1.80 and 1.931 values between 1.93 and 2.061 values between 2.06 and 2.201 values between 2.85 and 2.98offnegativepositiveoften

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. ransition
  2. 丰富
  3. REEN
  4. SG
  5. ОО
  6. 尽可能
  7. GR
  8. MERCHANTABILITY
  9. rendering
  10. reen

Suppresses

  1. ely
  2. ولا
  3. 的模样
  4. sometimes
  5. aldi
  6. yped
  7. ameda
  8. 的样子
  9. 唆
  10. iores