Neuron Explorer

Qwen3-0.6B

'for' after verbs like 'fall' or 'click'

A neuron in layer 4 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 responds to the word 'for' when it follows a verb, as in 'fallen for', 'stretched for', or 'Click for the answers'. Turning it up did not give clearly related text in the sample outputs.

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

    fossil that definitively prove their hunch, and are able to reveal the precise statistics of the giant herbivore. It seems that the beast would have stretched for 20 meters

  2. #2 peak

    PDF [portable document format] printable page. (If you are unable to view the printable page, download the free Adobe Acrobat Reader.)⏎Click for the answers to this week's puzzle.⏎Click for Sudoku Puzzles from previous

  3. #3 peak

    then hawking their services as conduits to the afterlife to the credulous grieving public commonly used. (Can you imagine anyone being foolish enough to fall for such predatory charlatanry today?)⏎That turn in

  4. #4 peak

    itself as Bodoland Martyrs Cemetery. Inside the Martyrs Cemetery, several hundred blocks of black granite are inscribed with names and addresses of men who have fallen for Bodoland. Life-size granite statues of leaders of the Bodo movement and insurgency

  5. #5 peak

    They have been sunk into E. plukenetii and are recognized as subsp. pl<|endoftext|>| ||Minding Your Mind || |⏎When Striving for Perfection Is a Problem⏎Last Reviewed by Faculty of Harvard Medical School on May

  6. #6 peak

    family members and government officials to change the trajectory of Alzheimer's disease," concluded Geiger.⏎The research was slated for presentation Monday at the Alzheimer's Association International Conference in Paris. Research presented at medical

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 am a new user. Please don't use the general public. This is a personal website. I need to be a CEO, a CTO, a CTO, and a CTO. This is a CEO, a CTO

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

    some easy and age-appropriate science fair project ideas for kids that will keep your kid excited about learning.⏎1. Lava Lamp⏎Image Credit: funlearningforkids.com⏎With this project, you and

  2. #2 peak

    the Malian military from the north, however, the separatist forces soon became embroiled in a conflict of their own, between the original Mouvement National pour la Libération de l’

  3. #3 peak

    waste specialist consultancy, Eunomia, who confirm the credibility of APEAL’s methodology and data.⏎www.apeal.org and www.steelforpackaging.org<|endoftext|>Paint Analysis - Peeling Paint? How to Diagnose

  4. #4 peak

    these prepositions must be in accusative case: durch-throughfuer-for gegen-againstohne-without um-around Er kauft das Geschenk fuer mich.<|endoftext|>In case you are searching for easy Preschool Activities, you can

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. L3:381 excites 'for' in research institute names
  2. L1:1341 suppresses
  3. L2:75 weakly suppresses
  4. L0:1543 weakly excites
  5. L2:2328 weakly suppresses
  6. L3:1390 weakly excites 'that' in 'so that' purpose clauses
  7. L2:1860 weakly excites
  8. L3:691 weakly excites
  9. L0:2711 weakly suppresses
  10. L1:1464 weakly excites
  11. L2:41 weakly suppresses
  12. L2:49 weakly suppresses
  13. L2:549 weakly suppresses
  14. L2:190 weakly suppresses
  15. L1:1972 weakly suppresses
  16. L3:456 weakly suppresses
  17. L0:4 weakly suppresses
  18. L3:2766 weakly excites
  19. L3:464 weakly excites
  20. L3:2818 weakly excites
  21. L3:1164 weakly excites
  22. L1:1306 weakly excites
  23. L0:1301 weakly suppresses
  24. L2:527 weakly excites

Feeds later layers

  1. L9:2587 suppresses
  2. L17:1223 suppresses
  3. L19:1221 suppresses
  4. L5:1090 suppresses
  5. L16:2336 suppresses
  6. L9:1890 suppresses
  7. L10:132 suppresses
  8. L5:763 suppresses
  9. L7:1172 suppresses 'for' after past participles like selected
  10. L17:2949 suppresses 'for' introducing an example or reason
  11. L10:1029 suppresses
  12. L9:1550 suppresses 'to' after 'matters' or 'means' phrases
  13. L16:524 suppresses
  14. L21:2012 suppresses "to" before an infinitive or noun
  15. L5:2538 suppresses
  16. L17:697 suppresses
  17. L12:2212 suppresses 'to' or 'for' opening a purpose phrase
  18. L10:2412 weakly suppresses
  19. L6:2899 weakly suppresses
  20. L15:2573 weakly suppresses
  21. L17:2302 weakly suppresses 'in' after a noun
  22. L17:2527 weakly suppresses
  23. L5:892 weakly suppresses
  24. L5:1864 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.

2 values between -6.73 and -6.531 values between -5.48 and -5.271 values between -4.64 and -4.431 values between -4.01 and -3.801 values between -3.80 and -3.592 values between -3.59 and -3.382 values between -2.96 and -2.752 values between -2.54 and -2.333 values between -2.33 and -2.121 values between -2.12 and -1.912 values between -1.91 and -1.702 values between -1.70 and -1.491 values between -1.49 and -1.283 values between -1.28 and -1.072 values between -1.07 and -0.861 values between -0.86 and -0.659 values between -0.65 and -0.443 values between -0.44 and -0.23206 values between -0.23 and -0.023097 values between -0.02 and 0.19531 values between 0.19 and 0.4033 values between 0.40 and 0.611 values between 1.03 and 1.241 values between 1.45 and 1.66offnegativepositiveoften

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. 都会
  2. ammed
  3. auses
  4. 都会有
  5. 还是会
  6. 会在
  7. istediği
  8. 会有
  9. 会
  10. ernet

Suppresses

  1. bidden
  2. HCI
  3. lox
  4. .getColumnIndex
  5. @student
  6. دا
  7. lix
  8. @show
  9. blister
  10. bind