Neuron Explorer

Qwen3-0.6B

middle of technical terms in science 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 the piece of a scientific or technical name that comes after the start, such as 'one' in 'pheromone', 'ome' in 'proteasome', or 'Survey' in 'Geological Survey'. It is mostly about word completion in science writing. Turning it up made the model repeat simple sentences, and turning it down gave unrelated story text.

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

    only the males of the species got attracted towards the light source, mistaking it for female moths. Female moths secrete a chemical called pheromone, which has a scent and acts as sex appeal for moths. The infra

  2. #2 peak

    either can’t make enough insulin to keep blood sugar levels down (type 1), or can’t effectively use the little insulin that it does produce (type 2).⏎First, to understand what happens in the bodies of people who have diabetes,

  3. #3 peak

    of the gardenia’s foliage. They thrive when the weather is warm and dry. Whiteflies feed on the gardenia’s sap and secrete a sticky substance known as honeydew. This, in turn, attracts sooty mold,

  4. #4 peak

    products, have been detected in about one-third of the wells sampled by the National Water-Quality Assessment (NAWQA) Program of the U.S. Geological Survey (USGS).1 Chloroform and other trihalomethanes,

  5. #5 peak

    endonuclease from the 2009 pandemic influenza strain using the high intensity X-ray beams at the European Synchrotron Radiation Facility (ESRF).⏎The researchers were able to determine the 3D atomic structure of the enzyme

  6. #6 peak

    itin side chains onto whatever protein is in close proximity, and when a protein is thus ubiquitinated, it is marked for destruction by the cell’s proteasome.⏎This means that you can,

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 have been, so, and I will be, and I will be, and I will be, and I will be, and I will be, and I will be, and I will be, and I will be, and 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

    &D and Opus 12 Present: “Utilization of CO2 Emissions to Make Renewable Fuels and Chemicals"⏎Wednesday, July 29, 2020⏎10:00 AM PDT.⏎Opus

  2. #2 peak

    ; hence, the popular phrase "Proverbs 31 Woman," while ignoring the advice for men. However

  3. #3 peak

    SS⏎Using an SPSS data file, you'll test hypotheses with t-tests and regression analysis to predict risks and trends.⏎Tuesday, Oct. 14, 6-8 p.m.⏎If you are registered for a quantitative course at

  4. #4 peak

    the Abnormal Silverline (Cigaritis abnormis), the Giant Red Eye (Gangara thyrsis) and the Large

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. L9:994 excites
  2. L9:571 excites
  3. L5:2491 suppresses
  4. L8:634 excites
  5. L8:621 excites
  6. L8:1219 weakly suppresses
  7. L6:1180 weakly suppresses
  8. L7:1019 weakly suppresses
  9. L7:1924 weakly suppresses
  10. L8:2142 weakly suppresses
  11. L9:742 weakly suppresses "of" after an agency or department name
  12. L8:658 weakly suppresses
  13. L5:320 weakly suppresses
  14. L9:247 weakly excites
  15. L7:1821 weakly suppresses
  16. L7:931 weakly suppresses
  17. L9:1045 weakly suppresses
  18. L8:877 weakly excites
  19. L5:666 weakly suppresses
  20. L1:779 weakly excites
  21. L9:694 weakly suppresses
  22. L4:10 weakly excites
  23. L8:699 weakly excites
  24. L9:993 weakly excites

Feeds later layers

  1. L12:590 suppresses
  2. L15:2570 suppresses
  3. L12:2477 excites
  4. L11:399 suppresses
  5. L13:761 weakly excites
  6. L11:237 weakly suppresses
  7. L11:1025 weakly excites
  8. L11:1239 weakly suppresses
  9. L11:767 weakly suppresses
  10. L20:1684 weakly suppresses
  11. L22:847 weakly excites
  12. L15:1478 weakly suppresses
  13. L12:778 weakly suppresses
  14. L16:1016 weakly excites
  15. L12:143 weakly excites
  16. L15:669 weakly suppresses
  17. L13:385 weakly excites
  18. L18:2049 weakly excites
  19. L12:1825 weakly suppresses
  20. L16:1223 weakly suppresses
  21. L20:333 weakly suppresses
  22. L12:1092 weakly excites
  23. L12:1700 weakly suppresses
  24. L14:78 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 -1.95 and -1.822 values between -1.69 and -1.572 values between -1.57 and -1.444 values between -1.44 and -1.319 values between -1.31 and -1.187 values between -1.18 and -1.0524 values between -1.05 and -0.9224 values between -0.92 and -0.7930 values between -0.79 and -0.6664 values between -0.66 and -0.53112 values between -0.53 and -0.40159 values between -0.40 and -0.27338 values between -0.27 and -0.14964 values between -0.14 and -0.011244 values between -0.01 and 0.12452 values between 0.12 and 0.25184 values between 0.25 and 0.3895 values between 0.38 and 0.5161 values between 0.51 and 0.6434 values between 0.64 and 0.7726 values between 0.77 and 0.9033 values between 0.90 and 1.038 values between 1.03 and 1.164 values between 1.16 and 1.298 values between 1.29 and 1.422 values between 1.42 and 1.557 values between 1.55 and 1.684 values between 1.68 and 1.811 values between 1.94 and 2.071 values between 2.07 and 2.202 values between 2.20 and 2.331 values between 2.59 and 2.721 values between 3.11 and 3.24offnegativepositiveoften

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. olds
  2. sons
  3. AYS
  4. annis
  5. actics
  6. IELDS
  7. nger
  8. yssey
  9. ****
  10. 丝

Suppresses

  1. Hogan
  2. vetica
  3. ToDevice
  4. VELO
  5. 的路上
  6. PointerException
  7. DOCUMENT
  8. 有一次
  9. iqueta
  10. ukkit