Neuron Explorer

Qwen3-0.6B

answer labels and tag questions

A neuron in layer 17 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

This neuron responds most to the word 'Answer' or a colon right before a reply, as in study worksheets. It is strongly pushed the other way by tag questions such as 'isn't it?' or 'doesn't he?'. Turning it down made the model produce a stream of 'right?' tag questions.

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

    are parallel, perpendicular or neither?⏎Parallel lines are the lines that never intersect and perpendicular lines are the lines that intersect at 90 degrees.⏎Answer: Lines with the same slope are parallel and if the slope of one line is the

  2. #2 peak

    is the total probability of finding a particle in a one-dimensional box in level n = 3 between x = 0 and x = L/ 6?⏎Answer to relevant QuestionsWhich has the lowest (ground-state) energy, an electron trapped

  3. #3 peak

    70%.⏎Answer: B. Humidity levels should be kept below 60 percent, ideally between

  4. #4 peak

    she could read prior to enrolling in the course. She obtained the following five number summary:⏎Use this information to draw a boxplot of the reading speed.⏎Answer to relevant QuestionsThe following

  5. #5 peak

    done.⏎Answer: True. There are hundreds of thousands of dust mites in our homes,

  6. #6 peak

    's disease, and how do you differentiate between regular memory loss and memory loss associated with Alzheimer's disease?⏎Answer: So, the memory changes of Alzheimer's disease represent a change for that person

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: right, right, right, right, right, right, right, right, right, right, right, right, right, right, right, right, right, right, right, right, right, right, right, right,

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

    inventor, Stairs. It's funny how you walked the stairs so many times and had no idea about who invented them and why they are called that, isn't it?<|endoftext|>Conventional, high-yield farming could have a better shot at

  2. #2 peak

    (TLB)⏎Jesus delayed on purpose so God’s glory would be shown and others would believe in him.⏎Sometimes Jesus does the same with our requests, doesn’t he?⏎He delays. We know he can. No doubt. We pray he

  3. #3 peak

    is simpler for you to collaborate with, and will certainly use up much less of your time than searching for individual worksheets.⏎This sounds like a horrible lot, does not it! A great collection of preschool worksheets should cover every one of this, and

  4. #4 peak

    craft (or lolly) sticks and rubber bands.⏎An amazing STEM project for future scientists – sling shot LEGO car! It even sounds so much fun, don’t you think?⏎Setting up fine motor practice really doesn’t have to be hard or

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. L15:2412 suppresses
  2. L16:1059 excites
  3. L14:441 weakly excites
  4. L8:1298 weakly excites
  5. L16:741 weakly suppresses
  6. L4:2258 weakly suppresses
  7. L8:2275 weakly suppresses
  8. L9:846 weakly excites
  9. L15:1388 weakly excites
  10. L1:1983 weakly excites
  11. L15:348 weakly suppresses
  12. L12:44 weakly suppresses
  13. L4:2564 weakly suppresses
  14. L4:1922 weakly suppresses
  15. L1:1950 weakly suppresses
  16. L16:376 weakly excites 'or' joining alternatives
  17. L3:2668 weakly excites
  18. L9:2524 weakly excites
  19. L14:2026 weakly suppresses
  20. L0:918 weakly excites
  21. L2:760 weakly excites
  22. L7:2377 weakly excites
  23. L0:1077 weakly excites
  24. L13:417 weakly excites

Feeds later layers

  1. L20:2335 strongly suppresses comma ending an 'if' clause
  2. L24:2138 suppresses comma before 'right?' tag
  3. L18:594 excites
  4. L20:2730 suppresses
  5. L21:1538 suppresses
  6. L19:835 suppresses
  7. L20:1922 suppresses
  8. L24:402 suppresses
  9. L20:1957 weakly suppresses
  10. L26:613 weakly suppresses
  11. L21:1362 weakly suppresses
  12. L25:541 weakly suppresses
  13. L21:1484 weakly suppresses comma after a sentence subject
  14. L22:387 weakly excites
  15. L19:2793 weakly suppresses
  16. L21:1044 weakly suppresses
  17. L20:459 weakly excites
  18. L18:1912 weakly excites
  19. L20:1838 weakly excites
  20. L24:2161 weakly suppresses
  21. L24:2257 weakly suppresses
  22. L26:1937 weakly suppresses
  23. L19:670 weakly excites
  24. L20:1371 weakly suppresses comma inside a question

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 -4.22 and -4.061 values between -3.60 and -3.451 values between -3.29 and -3.141 values between -2.37 and -2.211 values between -1.75 and -1.601 values between -1.14 and -0.984 values between -0.98 and -0.837 values between -0.83 and -0.6724 values between -0.67 and -0.5286 values between -0.52 and -0.37292 values between -0.37 and -0.21964 values between -0.21 and -0.061357 values between -0.06 and 0.10726 values between 0.10 and 0.25284 values between 0.25 and 0.41106 values between 0.41 and 0.5632 values between 0.56 and 0.719 values between 0.71 and 0.877 values between 0.87 and 1.021 values between 1.02 and 1.171 values between 1.33 and 1.481 values between 1.48 and 1.641 values between 1.79 and 1.94offnegativepositiveoften

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. _ly
  2. blockbuster
  3. bled
  4. tridge
  5. 骘
  6. 感动
  7. _mode
  8. OMATIC
  9. erry
  10. ONY

Suppresses

  1. 吗
  2. 吧
  3. ?
  4. right
  5. 嗎
  6. ?⏎
  7. 了吗
  8. 么
  9. 右
  10. ですよね