Neuron Explorer

Qwen3-0.6B

science news body text

A neuron in layer 10 of Qwen3-0.6B. Its strongest snippets are mixed: they peak on different tokens. It is featured, so it has steering outputs.

What this neuron does

This neuron fires on ordinary words inside news stories about research and discoveries, such as 'research', 'most' or 'detailed'. It goes strongly negative on 'says' when a quote is attributed to a scientist. Turning it up or down did not change the writing in a clear, themed 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

    ) Jonghwan Suhr, an assistant professor of mechanical engineering, recently published his research findings on Continuous Reinforced Carbon Nanotube Composites in Nano Letters.⏎Su

  2. #2 peak

    at the University of Edinburgh.⏎DePolo was the lead writer of a report on the study. The findings were published in the journal PLOS ONE.⏎The most famous member of the stegosaurs was Stegosaurus, which lived late in

  3. #3 peak

    trapped in the resin, which later fossilised into amber.⏎The details of the discovery are published in the journal Naturwissenschaften. (ANI)⏎- Oldest mammalian hair found in France - May 20, 201

  4. #4 peak

    ke said.⏎Data combined from it and two additional stations — one at Ocean View Community Beach in Norfolk and another at Sunset Beach Resort on the Eastern Shore — assist the Coast Guard in defining search areas for lost and drifting boats.⏎The data also can

  5. #5 peak

    's Department of Biomedical Sciences and Pathobiology, her husband Jatinder Singh, a research assistant professor and molecular biologist at Virginia Tech, and their daughter have spent most of 2008 living in primitive facilities near the shores of

  6. #6 peak

    in the area and was probably a little bit marshy.”⏎He also added that the country's laws permit property owners to keep fossils discovered on their premises.⏎Talking to

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 think the study says that the company is looking to use the data from the study." This is the statement from the study, the company. So, the question is about the study, the company. The statement says: "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

    Earth's surface. But when researchers looked at the western foothills southeast of Fresno, "we were no longer able to identify where the Moho was present," says Hersh Gilbert, a co-author

  2. #2 peak

    rites are shorter and have fewer branches than those of controls. The shortage of branches may translate to fewer synapses, or neuronal junctions, LaMantia says.⏎The mitochondria in the mutant mice’s neurons are also misshapen

  3. #3 peak

    computer industry aims to progress. And Apple’s M1 Ultra shows that new techniques can produce big leaps in performance.⏎“This technology showed up at just the right time,” says Tim Millet, vice president of hardware technologies at Apple. “In a

  4. #4 peak

    .⏎Hitesh Arora, co-lead author of the study who conducted this research as a Ph.D. student at the Michael Smith Laboratories at UBC said, “Sepsis triggers an uncontrolled chain-reaction of inflammation in the

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. L9:520 suppresses
  2. L9:428 excites split word pieces in names and years
  3. L9:5 weakly excites
  4. L2:1373 weakly excites
  5. L7:1179 weakly excites
  6. L7:1165 weakly suppresses
  7. L1:1560 weakly excites
  8. L8:7 weakly excites
  9. L7:2492 weakly suppresses
  10. L9:1887 weakly excites
  11. L9:834 weakly suppresses
  12. L8:2883 weakly excites
  13. L9:748 weakly excites
  14. L9:2897 weakly excites
  15. L2:2648 weakly suppresses
  16. L1:491 weakly excites
  17. L6:2176 weakly excites
  18. L7:695 weakly excites
  19. L9:2519 weakly excites
  20. L8:472 weakly suppresses
  21. L9:258 weakly suppresses
  22. L7:614 weakly suppresses
  23. L8:1961 weakly excites
  24. L1:2597 weakly excites

Feeds later layers

  1. L17:2382 suppresses
  2. L12:2969 suppresses
  3. L12:2838 suppresses
  4. L20:1065 suppresses
  5. L11:2605 suppresses
  6. L20:3052 suppresses
  7. L21:450 suppresses
  8. L19:1409 suppresses
  9. L14:2218 excites
  10. L11:3037 suppresses
  11. L11:2328 suppresses
  12. L20:2766 weakly excites
  13. L14:879 weakly suppresses
  14. L22:2649 weakly suppresses
  15. L19:2488 weakly excites
  16. L27:1630 weakly excites
  17. L15:1003 weakly suppresses
  18. L18:2285 weakly suppresses
  19. L20:1580 weakly suppresses
  20. L15:1940 weakly suppresses
  21. L14:1677 weakly suppresses
  22. L11:2414 weakly suppresses
  23. L12:2165 weakly suppresses
  24. L11:2215 weakly suppresses closing quote marks at the end of speech

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 -12.25 and -11.911 values between -8.22 and -7.891 values between -6.88 and -6.541 values between -6.21 and -5.871 values between -4.87 and -4.532 values between -4.53 and -4.191 values between -4.19 and -3.863 values between -3.52 and -3.192 values between -3.19 and -2.854 values between -2.85 and -2.521 values between -2.52 and -2.184 values between -2.18 and -1.845 values between -1.84 and -1.513 values between -1.51 and -1.177 values between -1.17 and -0.8422 values between -0.84 and -0.50615 values between -0.50 and -0.172701 values between -0.17 and 0.17507 values between 0.17 and 0.5124 values between 0.51 and 0.842 values between 0.84 and 1.18offnegativepositiveoften

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. 风俗
  2. '/');⏎
  3. undy
  4. ()];⏎
  5. lessly
  6. JSGlobal
  7. IGATION
  8. '/';⏎
  9. 祖国
  10. Subview

Suppresses

  1. 她说
  2. said
  3. ices
  4. Big
  5. 他说
  6. "
  7. said
  8. _comment
  9. istr
  10. Talk