Neuron Explorer

Qwen3-0.6B

words after comparisons and contrasts

A neuron in layer 17 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 small connecting words like 'than', 'in', 'from' or 'the' that set up a comparison or contrast, such as 'cleaner than' or 'differs from'. Turning it up gave loose, unrelated text, so the steering results do not clearly reflect this.

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

    .⏎Quantum computing combines computer science with quantum physics. These super computers store their data using sub-atomic particles known as qubits, instead of silicone chips used in traditional machines. Silicon chips store data in binary code – a series of ones and

  2. #2 peak

    large distance from the Sun, 40 times further than the Earth. It is orbiting in a plane that is far away from the nearly common plane in which the other planets are orbiting (know as the ecliptic). Moreover,

  3. #3 peak

    of Edinburgh and Copenhagen to investigate if they could outperform conventional machines for specific applications.⏎They examined whether quantum computing’s noise rates would prevent the new systems outperforming classical computers on specific tasks.⏎Experts developed a technique to compare how classical algorithms and

  4. #4 peak

    ’s Mouth Cleaner Than a Human Mouth? The general answer to this question is no, a dog’s mouth is not technically clean, and is not actually any cleaner than a human mouth.⏎Is saliva of dog harmful?⏎Recent case reports demonstrate the harms

  5. #5 peak

    . The bugs look quite similar and have the same life cycle, nesting areas, and impact on humans.⏎The primary thing that differentiates the tropical bed bug from the common bed bug is their habitat. While the common<|endoftext|>448 pp.,

  6. #6 peak

    all the other stars. The nearest star to our solar system is Proxima Centauri. The light of that star takes 4.25 years to reach Earth. That’s how long it takes light to travel 40,2

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 "the" is the "the" is the "the..." is the "..." is the "..." is "..." is the "..." is "..." is "..." is "..." is "..." is "..." is "..." 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

    two major types of roots, fibrous roots and taproots. Taproots have one major root that grows very deep into the ground, while fibrous roots have many roots that grow in all directions.⏎Image: DonNichols/Getty Images⏎The

  2. #2 peak

    , exposing them to hazardous and toxic material. They face harassment and discrimination because of a negative perception of their work. Women and children are especially vulnerable to violence both within and outside their homes. Substance abuse is high and accessing affordable healthcare is a huge

  3. #3 peak

    64-bit or 128-bit encryption key that must be manually entered on wireless access points and devices and does not change. TKIP employs a per-packet key, meaning that it dynamically generates a new 128-bit

  4. #4 peak

    round-the-clock care that's given by professional health care providers such as nurses, therapists, or aides under a doctor's supervision.⏎- Intermediate care: Also provided by professional health care providers but on a less

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. L16:427 excites
  2. L16:2812 weakly excites
  3. L13:1598 weakly excites
  4. L12:1142 weakly suppresses
  5. L10:1933 weakly suppresses
  6. L10:1453 weakly suppresses
  7. L10:2837 weakly suppresses
  8. L7:2594 weakly excites
  9. L16:2763 weakly suppresses
  10. L1:2882 weakly suppresses
  11. L15:1338 weakly suppresses
  12. L4:1760 weakly suppresses
  13. L10:2939 weakly excites
  14. L11:1288 weakly suppresses
  15. L12:1726 weakly suppresses
  16. L15:1869 weakly excites
  17. L11:2824 weakly suppresses
  18. L13:756 weakly suppresses
  19. L15:369 weakly excites
  20. L2:1259 weakly excites
  21. L2:1754 weakly excites
  22. L7:1261 weakly suppresses
  23. L0:1583 weakly suppresses
  24. L15:2613 weakly suppresses

Feeds later layers

  1. L18:439 strongly excites
  2. L18:257 strongly excites
  3. L18:206 strongly excites
  4. L18:667 strongly suppresses
  5. L18:2662 strongly suppresses
  6. L18:1088 strongly suppresses 'and' or 'the' joining two parallel items
  7. L19:522 suppresses
  8. L19:666 suppresses
  9. L20:165 excites
  10. L20:524 suppresses
  11. L18:362 suppresses
  12. L18:438 suppresses
  13. L18:255 suppresses
  14. L18:449 suppresses
  15. L19:2484 suppresses
  16. L18:301 suppresses
  17. L19:2897 suppresses
  18. L19:2298 suppresses
  19. L18:213 suppresses
  20. L19:519 suppresses
  21. L18:684 suppresses
  22. L20:294 excites
  23. L18:2953 suppresses
  24. L19:631 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 -4.88 and -4.641 values between -4.40 and -4.161 values between -3.92 and -3.681 values between -3.68 and -3.443 values between -3.44 and -3.2013 values between -3.20 and -2.9614 values between -2.96 and -2.7221 values between -2.72 and -2.4830 values between -2.48 and -2.2429 values between -2.24 and -2.0056 values between -2.00 and -1.7675 values between -1.76 and -1.5292 values between -1.52 and -1.28165 values between -1.28 and -1.04267 values between -1.04 and -0.80332 values between -0.80 and -0.57528 values between -0.57 and -0.33669 values between -0.33 and -0.09729 values between -0.09 and 0.15400 values between 0.15 and 0.39194 values between 0.39 and 0.63114 values between 0.63 and 0.8773 values between 0.87 and 1.1132 values between 1.11 and 1.3517 values between 1.35 and 1.5912 values between 1.59 and 1.836 values between 1.83 and 2.079 values between 2.07 and 2.317 values between 2.31 and 2.555 values between 2.55 and 2.792 values between 2.79 and 3.033 values between 3.03 and 3.271 values between 3.27 and 3.513 values between 3.51 and 3.751 values between 3.75 and 3.991 values between 4.46 and 4.70offnegativepositiveoften

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. ched
  2. itional
  3. zenie
  4. Williamson
  5. строй
  6. lement
  7. zburg
  8. exterity
  9. chts
  10. pection

Suppresses

  1. timing
  2. 稠
  3. 腊
  4. aure
  5. lique
  6. ủng
  7. uja
  8. sausage
  9. sequ
  10. oric