Neuron Explorer

Qwen3-0.6B

"that" starting a relative clause

A neuron in layer 21 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

The neuron lights up on the word "that" when it introduces a describing clause after a noun, like "the parts of the brain that control breathing." When it was turned up, the model fell into repeating "is" over and over.

Written by Claude from this neuron's activations and steering outputs. Confidence: high. 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

    the decimal and cents to the right of the decimal. To add decimals, simply line up the decimal points of the amounts of money that are being added. When you have added the amounts of money, be sure to

  2. #2 peak

    Center expect these cancer vaccines first will be used as supplements to initial treatment plans that use surgery, chemotherapy or radiation. The vaccine then will sweep out any cancer cells that remain in the body.⏎“Hypothetically, once the immune system has

  3. #3 peak

    not need an exact number, you can⏎round a number to make things simple.⏎Here are rules for rounding numbers⏎to the nearest ten.⏎Make the numbers that end in 1⏎through 4 into the next lower number that ends in

  4. #4 peak

    the herniation occurrs. Without treatment, death is likely.⏎A brain herniation often causes a massive stroke. There can be damage to parts of the brain that control breathing and blood flow. This can rapidly lead to death or brain death.⏎

  5. #5 peak

    before you start adding any additional chemicals to ensure active chlorine production is optimised.⏎- Put all your pool equipment e.g. hoses, poles and any other accessories that may be infected, in the shallowest part of your pool so that when you

  6. #6 peak

    a scarce resource,” he says. “So, by using the wastewater you’re really repurposing something that you’d otherwise pay to be processed and converting<|endoftext|>Career and technical education courses utilize

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: What if the new government which which which which which which which which which which which which which which which which which which which which which which which which which which which which which which which which which which which which which which which which which which which

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

    problem?⏎Our tutors have indicated that to solve this problem you will need to apply the Conjugate Acids and Bases concept. If you

  2. #2 peak

    gone up since then. Then they adjust the gross spending numbers by this (“deflator”). This is good for the government because it is lower than what we believe to be the actual price inflation rate and makes GDP look better than it really is.

  3. #3 peak

    ervention. It therefore avoids wastage of resources. The objective was to compare the number of CT-scans done for minor head injury as compared to the number that would have been done if the CCHR was applied. Methods:

  4. #4 peak

    ruins of both these temple complexes cover a considerable area and are still very impressive. Nothing remains, however, of the houses, markets, palaces and gardens that must have surrounded the temples in ancient times. The principal feature in Egyptian social centers,

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. L17:1361 suppresses 'that' after verbs like recommend
  2. L18:960 excites
  3. L19:1703 excites
  4. L20:1000 suppresses
  5. L18:491 excites
  6. L18:2031 suppresses
  7. L20:349 excites
  8. L18:1389 excites
  9. L16:2975 suppresses
  10. L19:622 suppresses
  11. L17:227 excites
  12. L18:2787 excites
  13. L19:547 weakly suppresses
  14. L16:948 weakly suppresses
  15. L18:1513 weakly excites
  16. L20:636 weakly suppresses
  17. L17:137 weakly excites
  18. L19:1115 weakly suppresses
  19. L9:865 weakly suppresses
  20. L15:1834 weakly excites
  21. L8:199 weakly excites
  22. L2:205 weakly suppresses
  23. L13:344 weakly suppresses
  24. L19:1175 weakly suppresses

Feeds later layers

  1. L23:30 strongly suppresses
  2. L22:228 strongly excites
  3. L23:127 strongly suppresses
  4. L22:479 strongly excites
  5. L23:320 excites
  6. L25:9 suppresses
  7. L26:669 excites
  8. L22:518 excites
  9. L26:53 suppresses
  10. L23:1013 excites
  11. L26:318 excites
  12. L26:467 excites
  13. L27:604 excites
  14. L25:698 excites
  15. L27:804 excites
  16. L25:586 excites
  17. L27:321 excites
  18. L23:220 excites
  19. L24:241 suppresses
  20. L27:382 excites
  21. L24:252 excites
  22. L27:565 excites
  23. L24:1439 suppresses
  24. L27:461 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 -4.04 and -2.42332 values between -2.42 and -0.803496 values between -0.80 and 0.8223 values between 0.82 and 2.449 values between 2.44 and 4.066 values between 4.06 and 5.681 values between 5.68 and 7.302 values between 7.30 and 8.921 values between 10.53 and 12.152 values between 12.15 and 13.773 values between 13.77 and 15.391 values between 15.39 and 17.013 values between 17.01 and 18.632 values between 18.63 and 20.251 values between 20.25 and 21.872 values between 21.87 and 23.494 values between 23.49 and 25.102 values between 26.72 and 28.341 values between 28.34 and 29.961 values between 31.58 and 33.202 values between 33.20 and 34.822 values between 34.82 and 36.443 values between 38.06 and 39.672 values between 41.29 and 42.911 values between 42.91 and 44.531 values between 46.15 and 47.771 values between 47.77 and 49.391 values between 54.24 and 55.862 values between 59.10 and 60.72offnegativepositiveoften

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. is
  2. are
  3. could
  4. was
  5. may
  6. has
  7. 是一
  8. can
  9. includes
  10. were

Suppresses

  1. которое
  2. which
  3. whom
  4. Which
  5. Which
  6. which
  7. velopment
  8. who
  9. 哪一个
  10. 哪