Neuron Explorer

Qwen3-0.6B

"of" after minority, tells or news

A neuron in layer 0 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 "of" when it comes after words like "minority", "tells", "news" or "unsure", as in "a minority of children" or "tells of a time". It is negative on "of" after job titles such as "Executive Director of". Steering at higher strength gave rambling text that did not clearly match 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

    view of New York politics a century and more back, and it has me thinking how much worse things are now.⏎Golway’s book on Tammany tells of a time when favors and jobs were bought and paid for with votes and tith

  2. #2 peak

    healthy screen time use.⏎"The findings of this meta-analysis indicate that only a minority of children 5 years and younger are meeting screen time guidelines," the study pointed out

  3. #3 peak

    from a contaminated environment.⏎Chronic illness only occurs in those who are genetically susceptible to improper toxin clearance (about 25% of the population). This minority of people typically have genes that prevent their bodies from being able to “tag” the

  4. #4 peak

    of his life. At age thirty-seven, Jeanne Wakatsuki Houston recalls life at Manzanar through the eyes of the child she was. She tells of her fear, confusion, and bewilderment as well

  5. #5 peak

    explain why many young people have not responded so enthusiastically to the “market signals” about college. On average, college pays off, but for a sizeable minority of college graduates, the arithmetic is not so convincing.⏎Imagine that you are a recent

  6. #6 peak

    that can help inform adaptation support strategies of governments and non-government actors.⏎The report makes three main points. First, Respondents in both Ethiopia and Mali tell of changing patterns of risks, and the emergence of new ones. There is a clear

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: It's a bit more about the things I have been thinking about lately. So the thing I have been thinking about lately? I have been thinking about the things I have been thinking about lately? I have been thinking about the things I

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

    dam on the Yangtze River caused a “significant” increase in seismic activity along the dam’s reservoir.⏎Patricia Adams is an economist and the Executive Director of Probe International, an independent think-tank and watchdog over the environmental consequences of Canadian government

  2. #2 peak

    the Willowbrook Wildlife center on how to prevent and deter skunks from your property humanely. Any⏎resident in need of additional assistance should contact the Assistant Director of Village Services, to have the Village inspect their property for additional guidance and options.

  3. #3 peak

    “Who shouldn’t be interested in the environment? It is important to all of us; we only have one earth,” says James Eyring, Assistant Director of the Environmental Center, who is leading the Birds of Prey presentation, fitness trail

  4. #4 peak

    /AIDS requires an increased amount of investment. These investments need to be amplified in both sexual education and innovative HIV/AIDS prevention programs.⏎Both the Executive Director of UNAIDS and the Director-General of the World Health<|endoftext|>S: By:

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

None: this is the first layer.

Feeds later layers

  1. L1:32 excites
  2. L26:848 weakly excites
  3. L16:1840 weakly suppresses
  4. L17:134 weakly excites
  5. L23:2360 weakly excites
  6. L1:1179 weakly suppresses 'in' after 'result' (results in)
  7. L26:938 weakly excites
  8. L24:2816 weakly excites
  9. L25:2324 weakly suppresses
  10. L27:1149 weakly excites
  11. L18:420 weakly excites
  12. L22:961 weakly excites
  13. L17:637 weakly excites comma opening an interrupting phrase
  14. L18:875 weakly excites
  15. L2:1005 weakly suppresses the '6' in years like 196x
  16. L25:527 weakly suppresses
  17. L9:852 weakly suppresses
  18. L18:2333 weakly excites
  19. L1:21 weakly suppresses
  20. L3:1730 weakly excites
  21. L27:2731 weakly excites
  22. L27:1111 weakly excites
  23. L26:713 weakly suppresses
  24. L18:495 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.

2 values between -1.02 and -0.971 values between -0.97 and -0.921 values between -0.87 and -0.821 values between -0.82 and -0.772 values between -0.77 and -0.721 values between -0.61 and -0.563 values between -0.56 and -0.512 values between -0.51 and -0.463 values between -0.46 and -0.412 values between -0.41 and -0.366 values between -0.36 and -0.313 values between -0.31 and -0.2620 values between -0.26 and -0.2067 values between -0.20 and -0.15133 values between -0.15 and -0.10326 values between -0.10 and -0.051494 values between -0.05 and 0.001352 values between 0.00 and 0.05285 values between 0.05 and 0.10127 values between 0.10 and 0.1538 values between 0.15 and 0.2010 values between 0.20 and 0.2611 values between 0.26 and 0.317 values between 0.31 and 0.363 values between 0.36 and 0.411 values between 0.41 and 0.464 values between 0.46 and 0.511 values between 0.67 and 0.721 values between 0.72 and 0.771 values between 0.97 and 1.02offnegativepositiveoften

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. ąż
  2. natives
  3. numeral
  4. asionally
  5. ivated
  6. motion
  7. motions
  8. ARISING
  9. docs
  10. uffle

Suppresses

  1. lico
  2. dT
  3. rix
  4. lık
  5. enger
  6. enus
  7. 落实
  8. dde
  9. plode
  10. psc