Neuron Explorer

Qwen3-0.6B

opening bracket after a unit conversion

A neuron in layer 20 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 reacts to the opening bracket that comes right after a measurement, where the writer gives the same amount in another unit, like '5 in. (12.7 cm)'. Turning it up didn't make the replies about measurements, so the link shows only in the text it reads.

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

    The Central Business District as seen from the top of the Carlton Centre⏎|Municipality||City of Johannesburg|⏎|• Total||1.09 km2 (0.42 sq mi)|⏎|• Density||13,0

  2. #2 peak

    than 30 minutes after shutdown⏎- Use clean, sealed bottles and flush<|endoftext|>The capture envelope for this system is approximated by a 5-in. (12.7-cm) cube.⏎A prototype system for docking a miniature

  3. #3 peak

    docking port. The capture envelope for this system is approximated by a 5-in. (12.7-cm) cube centered on the front of the docking-port

  4. #4 peak

    bps. The net result of these simultaneous forces is a 25% increase in monthly transit fees, from $16,000 per month to $20,000 per month. The percentages vary

  5. #5 peak

    Visayan warty pigs seasonally grow large manes that start at the top of their head and travel partially down their back and grow up to 9 inches (23 centimeters) long! The mama of the group is Kurit (

  6. #6 peak

    which caused the deadliest tsunami in modern history – had such an effect.⏎Gross calculated that the Chilean quake shifted the Earth's figure axis by 3 inches (8 cm or 27 milliarcseconds), and shortened the length of

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 [map] is a [map] in the [map [map [map [map [map [map [map [map [map [map [map [map [map [map [map [map [map [map [

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

    Germs and other things that cause foodborne illness usually can be killed when heated. Heating food to an internal temperature above 160 degrees Fahrenheit or 78 degrees Celsius, even for a few seconds, is enough to kill most bacteria

  2. #2 peak

    3.0°F (22.8°C). The warmest month, on average, is November with an average temperature of 81.0°F (27.2°C). The coolest month on average is June, with an

  3. #3 peak

    inches of rainfall in 2016, the temperature in this arid region varies from freezing cold in winter to over 100º F (38º C) in summer.⏎It was also home to some of the oldest

  4. #4 peak

    6 25.........⏎Basics: Divide integers to get integer outputs in all cases⏎Determine value using Look-up table method⏎Create Volcano (or Atoll) martix. It is an extension of Bullseye matrix problem

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. L3:1911 excites
  2. L12:1332 weakly suppresses
  3. L10:1950 weakly excites
  4. L4:2313 weakly suppresses
  5. L9:1204 weakly suppresses
  6. L3:1312 weakly suppresses
  7. L6:2393 weakly suppresses
  8. L5:1704 weakly excites
  9. L13:2541 weakly suppresses
  10. L5:27 weakly excites
  11. L10:157 weakly suppresses end of document and web boilerplate
  12. L7:2440 weakly excites
  13. L17:1506 weakly suppresses
  14. L3:594 weakly suppresses
  15. L13:2186 weakly excites
  16. L1:779 weakly excites
  17. L5:86 weakly suppresses
  18. L6:1390 weakly excites
  19. L9:318 weakly suppresses
  20. L13:1723 weakly suppresses
  21. L9:2331 weakly excites
  22. L12:2617 weakly suppresses
  23. L17:172 weakly excites
  24. L17:1582 weakly suppresses

Feeds later layers

  1. L22:2388 suppresses
  2. L22:2840 suppresses 'as' in 'as low as' and 'as far as'
  3. L23:2845 weakly suppresses
  4. L21:2323 weakly suppresses
  5. L21:1848 weakly suppresses
  6. L21:2151 weakly excites
  7. L21:2637 weakly excites
  8. L23:1035 weakly suppresses
  9. L22:1671 weakly excites
  10. L25:1747 weakly suppresses
  11. L27:2916 weakly suppresses
  12. L21:995 weakly suppresses
  13. L24:2631 weakly suppresses
  14. L22:2747 weakly suppresses
  15. L27:2006 weakly suppresses
  16. L22:2360 weakly excites
  17. L27:2885 weakly suppresses
  18. L25:2292 weakly excites
  19. L27:1344 weakly suppresses
  20. L21:520 weakly excites
  21. L23:861 weakly excites
  22. L25:1353 weakly excites "in" before "good" condition
  23. L26:2143 weakly excites
  24. L22:2678 weakly 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.

1 values between -2.82 and -2.581 values between -1.85 and -1.601 values between -1.60 and -1.366 values between -1.36 and -1.122 values between -1.12 and -0.8739 values between -0.87 and -0.63159 values between -0.63 and -0.39758 values between -0.39 and -0.142551 values between -0.14 and 0.10320 values between 0.10 and 0.3443 values between 0.34 and 0.5912 values between 0.59 and 0.832 values between 0.83 and 1.074 values between 1.07 and 1.322 values between 1.56 and 1.802 values between 1.80 and 2.052 values between 2.05 and 2.291 values between 2.78 and 3.021 values between 3.26 and 3.511 values between 6.67 and 6.91offnegativepositiveoften

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. ائية
  2. ски
  3. emia
  4. presence
  5. 疡
  6. presence
  7. 祓
  8. /gin
  9. ском
  10. {name

Suppresses

  1. map
  2. maps
  3. map
  4. .map
  5. -map
  6. (map
  7. →map
  8. Map
  9. apply
  10. maps