Neuron Explorer

Qwen3-0.6B

'that' after reporting verbs

A neuron in layer 1 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 'that' when it follows a verb of saying or showing, such as 'asserts that', 'notes that' or 'showed that'. Turning it up did not make the model's writing obviously follow 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

    5-16, 1972. Representatives of 113 governments attended and adopted the Stockholm Declaration on the Human Environment, which asserts that humans have a fundamental right to an environment that permits a life of dignity and well

  2. #2 peak

    in there” to “restore fusion actions to [their] former glory.”⏎Supersymmetry and Q-balls are thus far only theoretical, although Cox asserts that he and the physicists at CERN will prove or disprove their existence by

  3. #3 peak

    ists contend that no matter what the amount used in dental fillings, the mercury is toxic. At the same time, the American Dental Association (ADA) asserts that only a small percentage of individuals exhibit allergies to mercury, and that once mercury is

  4. #4 peak

    these measurements are used to help identify the species of bat.⏎However, a paper by an international team of researchers, published in the Journal of Applied Ecology, asserts that poor standardisation of acoustic monitoring limits scientists’ ability to collate data.⏎Kate

  5. #5 peak

    in comparing the post-tsunami scenario with the pre-tsunami situation, particularly with regard to coastal landforms and characteristics of soil and water.⏎Noting that the studies included Madurai, Coimbatore, Salem and Pondicherry

  6. #6 peak

    , which diminishes a community’s ability to tackle complex problems, including poverty, as well as its ability to generate high-quality economic development.⏎The study also notes that higher poverty rates translate into higher costs for taxpayer-funded public assistance programs.⏎- Download

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 of the book, "The Art of Computer Programming, by Marko's " -" -" -" -" -" -" -" -" -" -" -" -" -" -" -

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

    boline alkaloids (which are monoamine oxidase inhibitors, or MAOIs) with N, N-Dimethyltryptamine (DMT) that produce the hallucinogenic effects of the brew.⏎Some say that the value of

  2. #2 peak

    2009, strains of K. pneumoniae with gene called New Delhi metallo-beta-lactamase (shortened NDM-1) that even gives resistance against intravenous antibiotic carbapenem, were discovered in India

  3. #3 peak

    MET protein.⏎MET is a receptor tyrosine kinase (RTK) that is produced as a single-chain precursor. The precursor is proteolytically cle

  4. #4 peak

    cheadlines⏎Google has started a new page on Google Plus to share their vision of

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. L0:1224 suppresses
  2. L0:178 suppresses
  3. L0:1301 weakly suppresses
  4. L0:2204 weakly excites
  5. L0:502 weakly excites
  6. L0:36 weakly suppresses
  7. L0:40 weakly excites 'of' in 'sort of' hedges
  8. L0:1311 weakly suppresses
  9. L0:302 weakly suppresses
  10. L0:453 weakly suppresses
  11. L0:2528 weakly excites
  12. L0:1764 weakly suppresses
  13. L0:1837 weakly excites
  14. L0:2278 weakly suppresses
  15. L0:277 weakly suppresses
  16. L0:2027 weakly suppresses
  17. L0:2230 weakly suppresses
  18. L0:729 weakly excites 'st' in dates like 'June 1st'
  19. L0:1025 weakly excites
  20. L0:612 weakly excites
  21. L0:386 weakly suppresses
  22. L0:510 weakly suppresses
  23. L0:879 weakly excites
  24. L0:2475 weakly suppresses

Feeds later layers

  1. L3:379 excites 'that' after 'the good news is'
  2. L4:227 excites
  3. L21:1606 weakly suppresses
  4. L2:505 weakly excites
  5. L4:223 weakly excites
  6. L11:1827 weakly excites
  7. L6:918 weakly excites
  8. L15:1020 weakly excites
  9. L10:2554 weakly suppresses
  10. L25:1364 weakly excites
  11. L20:1715 weakly excites
  12. L10:63 weakly suppresses
  13. L11:2031 weakly excites
  14. L18:2104 weakly suppresses
  15. L17:150 weakly excites
  16. L4:2757 weakly excites
  17. L19:1177 weakly excites
  18. L9:2817 weakly suppresses
  19. L18:2902 weakly excites
  20. L7:1833 weakly suppresses
  21. L15:1792 weakly suppresses
  22. L11:1723 weakly suppresses
  23. L20:533 weakly excites
  24. L15:1345 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 -1.49 and -1.321 values between -1.15 and -0.981 values between -0.98 and -0.812 values between -0.81 and -0.641 values between -0.64 and -0.4727 values between -0.47 and -0.30327 values between -0.30 and -0.132537 values between -0.13 and 0.04950 values between 0.04 and 0.2043 values between 0.20 and 0.371 values between 0.37 and 0.541 values between 0.54 and 0.712 values between 0.88 and 1.051 values between 1.05 and 1.221 values between 1.39 and 1.561 values between 1.56 and 1.731 values between 1.73 and 1.901 values between 2.24 and 2.411 values between 2.75 and 2.912 values between 2.91 and 3.081 values between 3.25 and 3.421 values between 4.10 and 4.272 values between 4.44 and 4.611 values between 4.61 and 4.781 values between 5.12 and 5.29offnegativepositiveoften

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. ônica
  2. ónica
  3. brace
  4. ónico
  5. onica
  6. awn
  7. rego
  8. ereal
  9. Alonso
  10. arus

Suppresses

  1. .SelectCommand
  2. 的表现
  3. held
  4. 颊
  5. 流逝
  6. cố
  7. 的地
  8. 的学习
  9. flap
  10. fet