Neuron Explorer

Qwen3-0.6B

comma after a place name

A neuron in layer 7 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 to commas that separate a city from its country or region, such as 'Sydney, Australia'. Turning it up did not produce any obvious change in the topic or style of the writing.

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 European Free Trade Association, Chile, Thailand, Costa Rica, Panama, Venezuela, Honduras, concluded negotiations with Guatemala and

  2. #2 peak

    Utilization and Innovation of Technology, 1 . pp. 30-35. ISSN 983-42733-2-3⏎This paper is a concept paper that will mainly discuss the on idea of

  3. #3 peak

    Gary White of Salisaw Equine Clinic, Salisaw, Okla. “I’m not sure whether it’s the confinement itself or just

  4. #4 peak

    Visit Event Page HERE!<|endoftext|>NORD is very grateful to M.Cristina Digilio, MD, Medical Genetics, Bambino Gesu Hospital, Rome, Italy, for assistance in the preparation of this report.⏎Syn

  5. #5 peak

    at Macquarie University, Sydney, Australia.⏎Browse the Lab Book to learn about the history, methods⏎and applications

  6. #6 peak

    km (11,100 miles) from the blast. These stations monitor seismic waves, infrasound (low frequency soundwaves), hydroacoustic, and radionuclide emissions as part of the Comprehensive Nuclear-Test-Ban

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 times that I have been thinking about lately. Let $x$ be real numbers such that $p, q, r \in \mathbb{R} $ are the product of the numbers are $p, q, r $

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

    cut to the length of the object you are working on.⏎put as many folds into the ruler as you can until it gets confusing: fourths/eighths/sixteenths, thirds/sixths, square root of 3 or

  2. #2 peak

    that many jumping spiders, crab spiders, wolf spiders and other garden spiders are not indiscriminate diners, but in fact

  3. #3 peak

    the whole comes before the individual’ and the direction of .all members of society by a single will supposed to represent the ‘whole’.<|endoftext|>Ears, Nose, and Throat Surgery⏎One of the most critical areas on the human body is

  4. #4 peak

    anode bar, and spacing distances between cathode and anode of 2, 3, and 4 in. (51, 76, and 102 mm). The specimens were maintained at a constant temperature and

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. L5:419 excites
  2. L5:159 weakly suppresses
  3. L5:531 weakly suppresses
  4. L6:533 weakly suppresses
  5. L4:182 weakly suppresses
  6. L4:321 weakly excites
  7. L4:273 weakly excites
  8. L4:502 weakly excites
  9. L2:298 weakly excites
  10. L4:436 weakly excites
  11. L4:276 weakly excites
  12. L5:216 weakly suppresses
  13. L6:136 weakly suppresses
  14. L5:2847 weakly excites
  15. L5:1435 weakly excites
  16. L6:2947 weakly suppresses
  17. L6:906 weakly excites
  18. L3:621 weakly excites
  19. L5:743 weakly suppresses
  20. L5:1510 weakly excites
  21. L5:1388 weakly suppresses
  22. L1:812 weakly suppresses
  23. L3:1222 weakly excites
  24. L1:734 weakly excites

Feeds later layers

  1. L10:554 suppresses
  2. L8:150 suppresses
  3. L10:1901 suppresses
  4. L9:622 suppresses
  5. L13:932 suppresses second digit of a year range
  6. L8:719 suppresses
  7. L14:2459 suppresses
  8. L12:1205 suppresses
  9. L13:1138 suppresses
  10. L13:697 suppresses
  11. L12:1700 suppresses
  12. L10:1511 suppresses
  13. L11:2551 suppresses
  14. L17:1740 suppresses
  15. L10:1912 weakly suppresses
  16. L8:336 weakly suppresses
  17. L14:1335 weakly suppresses
  18. L12:1813 weakly excites
  19. L9:2353 weakly suppresses
  20. L10:2968 weakly suppresses
  21. L14:2647 weakly suppresses 'and' or 'or' in split hyphen compounds
  22. L11:1024 weakly suppresses
  23. L13:2476 weakly suppresses
  24. L11:1960 weakly 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 -3.94 and -3.811 values between -2.96 and -2.841 values between -2.48 and -2.361 values between -2.36 and -2.231 values between -1.99 and -1.871 values between -1.75 and -1.631 values between -1.26 and -1.147 values between -1.02 and -0.906 values between -0.78 and -0.6613 values between -0.66 and -0.5322 values between -0.53 and -0.41110 values between -0.41 and -0.29429 values between -0.29 and -0.17959 values between -0.17 and -0.051523 values between -0.05 and 0.07599 values between 0.07 and 0.20179 values between 0.20 and 0.3243 values between 0.32 and 0.443 values between 0.44 and 0.564 values between 0.56 and 0.683 values between 0.68 and 0.801 values between 0.80 and 0.93offnegativepositiveoften

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. amodel
  2. aturally
  3. qp
  4. 皋
  5. \db
  6. cular
  7. bersome
  8. aml
  9. AMB
  10. ucing

Suppresses

  1. and
  2. Dis
  3. 散
  4. Pre
  5. Dis
  6. 닥
  7. 和支持
  8. and
  9. 장을
  10. 급