Neuron Explorer

Qwen3-0.6B

"x" or "by" between two dimensions

A neuron in layer 8 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 fires strongly on the "x" or "by" in measurements like "20 x 17 cm", "30 x 30" or "2 miles wide by 6 miles long". It is one of the clearest neurons in this set: it reacts to the middle of a size or dimension description. Steering at higher strength produced number-heavy text in some cases, which fits loosely.

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

    how it deals with death. Published by Edition Patrick Frey (Zurich).⏎250 pages, 20 x 17 cm, softcover, Edition Patrick Frey (Zurich

  2. #2 peak

    For example, you may find a threaded rod measuring M8-1.25 x 1 Meter. The M8 (8mm) is the rod's diameter

  3. #3 peak

    future town. It will be massive: 2 miles wide by 6 miles long and up to 10,000 homes.

  4. #4 peak

    ("Nanoholes As Nanochannels: Flow-through Plasmonic Sensing"), researchers at the University of British Columbia created a grid of 30 x 30 flow-through nanoscale holes to create a highly responsive sensor system

  5. #5 peak

    Northumberland).⏎The fort itself covers an area of about 4½ acres (c.1.8 ha), measuring 515 feet north-south by 385 feet east-west (157 x 117

  6. #6 peak

    , 1649.⏎Etching and engraving on laid paper, with margins; letterpress text on verso; platemark: 422 x 531 mm; total: 518 x 63

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: is the problem, or? I am, or? I, or? I, or? I, or? I, or? I, or? I, or? I, or? I, or? I, or? 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

    not died? (9) They had returned the home? (10)You had finished this book?⏎Past Perfect Tense Negative⏎Subject + had not + verb 3rd from + object.⏎Past Perfect Interrogative Negative (Examples

  2. #2 peak

    is expected to increase, which is a growing public health concern.⏎|Status||Published - 2014 jan. 1|⏎- Folkhälsovet

  3. #3 peak

    door size⏎Place the fully charged battery in an environment of 55°C±5°C. After the surface temperature of the battery reaches 55°C±5°C, leave it for 30 minutes, then connect the positive and

  4. #4 peak

    However, since it is a function instead of syntax, it can be partially applied and passed around to higher order functions, like so:⏎ghci> :m + Data.Bool.Extras⏎ghci> let yesOrNo = bool

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. L4:799 suppresses
  2. L4:760 excites
  3. L5:2079 excites
  4. L7:42 excites
  5. L7:1000 weakly excites
  6. L5:640 weakly suppresses
  7. L6:968 weakly suppresses
  8. L7:1664 weakly excites
  9. L1:360 weakly suppresses
  10. L4:717 weakly excites
  11. L3:726 weakly suppresses
  12. L6:2682 weakly excites
  13. L6:1084 weakly suppresses
  14. L2:112 weakly excites
  15. L1:2120 weakly excites
  16. L2:2009 weakly suppresses
  17. L5:117 weakly excites
  18. L1:660 weakly excites 'with' in 'along with'
  19. L3:1668 weakly suppresses
  20. L7:2962 weakly suppresses
  21. L6:604 weakly suppresses
  22. L3:2031 weakly excites
  23. L7:2320 weakly suppresses
  24. L7:23 weakly suppresses

Feeds later layers

  1. L12:51 suppresses
  2. L9:878 excites
  3. L14:706 excites
  4. L16:1074 excites
  5. L18:1857 excites
  6. L9:651 suppresses
  7. L16:314 excites
  8. L9:326 excites
  9. L12:336 suppresses
  10. L12:2548 excites
  11. L18:1270 excites
  12. L18:870 excites
  13. L9:212 suppresses
  14. L16:1999 excites
  15. L19:1823 excites 'and' inside organisation or paired names
  16. L9:1047 excites
  17. L12:1095 excites
  18. L10:1898 excites
  19. L14:917 excites
  20. L13:27 suppresses
  21. L11:682 excites gap before a second number
  22. L15:1162 excites
  23. L14:1054 excites
  24. L10:272 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.

2 values between -0.75 and -0.6612 values between -0.66 and -0.5721 values between -0.57 and -0.4851 values between -0.48 and -0.39149 values between -0.39 and -0.29304 values between -0.29 and -0.20651 values between -0.20 and -0.111392 values between -0.11 and -0.021130 values between -0.02 and 0.07137 values between 0.07 and 0.1636 values between 0.16 and 0.258 values between 0.25 and 0.344 values between 0.34 and 0.431 values between 0.43 and 0.522 values between 0.52 and 0.611 values between 1.07 and 1.163 values between 1.34 and 1.431 values between 1.88 and 1.971 values between 2.33 and 2.421 values between 2.70 and 2.791 values between 2.79 and 2.88offnegativepositiveoften

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. eten
  2. >'+⏎
  3. ">'+
  4. Department
  5. ">'+⏎
  6. <?⏎
  7. >"+⏎
  8. "").
  9. ultureInfo
  10. "\⏎

Suppresses

  1. umbai
  2. 滂
  3. amble
  4. like
  5. ович
  6. -know
  7. gressive
  8. wrapper
  9. /auto
  10. egade