Neuron Explorer

Qwen3-0.6B

medical tests, scans and assessments

A neuron in layer 11 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 words about medical checks and scanning, such as 'tests' before a kidney check, 'ultrasound' in a list of instruments, or 'assessment' in a home energy audit. Turning it up did not produce obviously medical text, so the theme is only partly clear.

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

    to brush crowded or crooked teeth. If your teeth don't get clean, they can get cavities and other problems.⏎There are special dentists who can tell if you need braces and can fit you with them if you do. They are

  2. #2 peak

    LIMITED WARRANTY*⏎All spray foam installations are carried out by Lapolla / AirTight licensed installers.⏎*See manufacturer guidelines and requirements.⏎HOME ENERGY ASSESSMENT⏎A home energy assessment helps owners determine where their house is losing energy and

  3. #3 peak

    , foam boxes, or 3D scanning⏎- Either vacuum forming thermoplastic over the cast or translating the scanned data into CAD/CAM processes, leading to subtractive milling⏎- Finishing

  4. #4 peak

    a baby. Along with placenta forceps, there are many other gynecological instruments which are used such as the vaginal speculum, retractors, and ultrasound machines and so on. These gynecological instruments help gynecologists to<|endoftext|>

  5. #5 peak

    - you need to have surgery in hospital⏎- you are trying to get pregnant, are pregnant or breastfeeding⏎Before prescribing lithium, your doctor will do some blood tests to check your kidney and thyroid are OK. The doctor

  6. #6 peak

    A related method is called “rational drug design,” where scientists “build” drug molecules specifically so they can fight HIV in targeted ways. Computers can be used to simulate a chemical compound and to design chemical structures that might work against HIV. While

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 thinking about the results of the results of the results... This is it. Now, the question is: What do you need to know about the results of the results of the results... This is it? After the question is

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

    baldness, the causes are also common and repeating. The most common factors that cause a receding hairline and male pattern baldness are:⏎1. Genetically Caused – Sometimes hair loss occurs because it runs in the family. Yes,

  2. #2 peak

    profound transformation of how energy is supplied and used around the world. The DDPP is charting the path to this transformation, starting now, one country at a time.⏎Deep decarbonization pathways fill a key gap in climate policy. Within countries

  3. #3 peak

    as a team.⏎It's important to discuss and agree on a sleep strategy for your child with your spouse or partner beforehand and work together as a team to carry it out consistently. Otherwise, you can't expect your child to learn or change her

  4. #4 peak

    day, you will give a synthetic (laboratory-made) form of thyroid hormone to your baby to replace the missing thyroid hormone. You’ll crush the hormone pill and mix it with a

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. L10:408 excites
  2. L8:220 suppresses
  3. L10:329 weakly suppresses
  4. L8:2613 weakly suppresses
  5. L10:573 weakly excites
  6. L9:512 weakly excites
  7. L4:2692 weakly excites
  8. L10:2763 weakly excites
  9. L3:1665 weakly excites
  10. L8:1597 weakly excites
  11. L7:1688 weakly excites
  12. L0:1889 weakly suppresses
  13. L10:412 weakly suppresses
  14. L5:631 weakly suppresses
  15. L4:1027 weakly excites
  16. L9:2450 weakly excites
  17. L10:20 weakly suppresses
  18. L1:1298 weakly suppresses
  19. L3:1870 weakly suppresses
  20. L6:2537 weakly excites
  21. L7:1803 weakly suppresses
  22. L8:867 weakly excites
  23. L6:1474 weakly suppresses
  24. L10:1853 weakly excites

Feeds later layers

  1. L15:1743 suppresses
  2. L18:1480 suppresses
  3. L16:1687 excites
  4. L12:1167 suppresses
  5. L15:1006 suppresses
  6. L13:1116 excites
  7. L16:2559 excites
  8. L16:2029 suppresses
  9. L14:1314 excites
  10. L22:1179 suppresses
  11. L17:789 suppresses
  12. L12:694 suppresses
  13. L12:1037 excites
  14. L15:2073 suppresses
  15. L13:363 suppresses
  16. L13:2677 excites
  17. L19:930 suppresses
  18. L15:595 suppresses
  19. L12:1953 suppresses
  20. L22:192 suppresses
  21. L12:1230 excites
  22. L17:2325 suppresses
  23. L14:537 excites
  24. L14:1822 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.

1 values between -2.34 and -2.241 values between -1.95 and -1.852 values between -1.85 and -1.752 values between -1.65 and -1.552 values between -1.55 and -1.452 values between -1.45 and -1.366 values between -1.36 and -1.263 values between -1.26 and -1.166 values between -1.16 and -1.067 values between -1.06 and -0.9616 values between -0.96 and -0.8720 values between -0.87 and -0.7727 values between -0.77 and -0.6734 values between -0.67 and -0.5746 values between -0.57 and -0.4768 values between -0.47 and -0.37130 values between -0.37 and -0.28215 values between -0.28 and -0.18485 values between -0.18 and -0.081626 values between -0.08 and 0.02793 values between 0.02 and 0.12201 values between 0.12 and 0.2282 values between 0.22 and 0.3144 values between 0.31 and 0.4125 values between 0.41 and 0.5118 values between 0.51 and 0.6110 values between 0.61 and 0.719 values between 0.71 and 0.804 values between 0.80 and 0.902 values between 0.90 and 1.007 values between 1.00 and 1.102 values between 1.10 and 1.206 values between 1.20 and 1.303 values between 1.30 and 1.391 values between 1.39 and 1.492 values between 1.49 and 1.59offnegativepositiveoften

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. uchs
  2. ycin
  3. ICommand
  4. %");⏎
  5. %"
  6. 和服务
  7. %"⏎
  8. Guidance
  9. astically
  10. went

Suppresses

  1. ?
  2. 测
  3. 測
  4. 测量
  5. measurement
  6. 检测
  7. 观测
  8. lf
  9. ometer
  10. 累