Neuron Explorer

Qwen3-0.6B

numbers inside journal citations

A neuron in layer 12 of Qwen3-0.6B. Its strongest snippets are mixed: they peak on different tokens. It is featured, so it has steering outputs.

What this neuron does

This neuron fires on the digits and journal names inside academic references, such as the volume or issue number in a citation. Turning it up made the model repeat phrases and go off topic.

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

    well represented in the efforts currently under way to elucidate the molecular basis of the genetic susceptibility to alcoholism.(JAMA. 1992;268:1877-1882)<|endoftext|>The Imagin

  2. #2 peak

    congestive heart failure;Malik et al., Cardiac Myosin Activation: A Potential Therapeutic Approach for Systolic Heart Failure. Science 331: 1439 (2011)). Comparative studies of

  3. #3 peak

    be able to survive in cities because if not, they will become extinct at the mercy of urban growth."⏎These results appear in the journal, Animal Behavior.<|endoftext|>NEW YORK – Researchers in Iceland led by Decode Genetics' Kari Stef

  4. #4 peak

    crystals, and could pave the way for all-optical circuits (M Mützel et al 2002 Phys. Rev. Lett. 88 083601-1).⏎The light field of an

  5. #5 peak

    has one of the highest rates of preterm birth, as well as density of Superfund Sites in the United States. As part of NIEHS's Superfund Research Program, the Puerto Rico Testsite for Exploring Contamination Threats (

  6. #6 peak

    4). You are what you can access: Sharing and collaborative consumption online. Journal of Business Research, 67(8), 1595-160

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: "X", and "Y", and "Z", and "Z", and "Z", and "Z", and "Z", and "Z", and "Z", and "Z", and "Z", and "Z", and

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 appear to be fearful. For example, aggression in response to ear cleaning or physical punishment might be examples of defensive aggression where the fear component is minimal. Fear aggression may be displayed when a dog is the recipient of assertive posturing or facial

  2. #2 peak

    a “Helper’s High”⏎In the 1980s, researchers coined a term called the “helper’s high.” It referred to the “high” that people felt when they helped others.

  3. #3 peak

    has not been understood completely, many people have remained skeptical," said Nedergaard, co-director of the University's Center for Translational Neuromedicine, where the research was conducted.⏎"In this work, we provide information about one

  4. #4 peak

    . Self-publishing, likewise called vanity publishing or self-publication, takes place when an author consents to allow a vanity

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. L11:1862 suppresses
  2. L11:325 suppresses
  3. L10:427 excites
  4. L9:247 excites
  5. L8:2187 weakly suppresses
  6. L10:1631 weakly suppresses
  7. L5:271 weakly excites
  8. L6:246 weakly excites
  9. L4:1185 weakly suppresses
  10. L8:242 weakly suppresses
  11. L9:828 weakly suppresses
  12. L11:838 weakly excites
  13. L8:2377 weakly excites
  14. L11:1888 weakly suppresses
  15. L8:949 weakly suppresses
  16. L6:1648 weakly excites
  17. L4:1861 weakly suppresses
  18. L10:2829 weakly suppresses
  19. L11:1990 weakly suppresses
  20. L6:987 weakly suppresses
  21. L7:2850 weakly excites
  22. L4:58 weakly suppresses
  23. L3:1547 weakly excites
  24. L10:892 weakly excites

Feeds later layers

  1. L13:1676 weakly suppresses
  2. L15:877 weakly suppresses
  3. L15:573 weakly suppresses
  4. L16:1883 weakly suppresses
  5. L13:271 weakly excites
  6. L18:265 weakly excites first half of institution names
  7. L24:1132 weakly suppresses
  8. L16:793 weakly excites
  9. L14:2564 weakly excites unclear: word before a technical term
  10. L21:2523 weakly suppresses
  11. L13:1345 weakly suppresses
  12. L13:1215 weakly suppresses
  13. L18:2574 weakly excites
  14. L24:596 weakly excites
  15. L23:290 weakly suppresses
  16. L15:1236 weakly suppresses
  17. L14:1807 weakly excites
  18. L15:2644 weakly suppresses
  19. L22:2168 weakly suppresses
  20. L24:2311 weakly excites
  21. L15:1128 weakly excites
  22. L18:2926 weakly excites
  23. L14:695 weakly excites
  24. L18:1226 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 -2.90 and -2.771 values between -2.38 and -2.242 values between -2.24 and -2.111 values between -2.11 and -1.982 values between -1.85 and -1.721 values between -1.72 and -1.592 values between -1.59 and -1.462 values between -1.46 and -1.327 values between -1.32 and -1.196 values between -1.19 and -1.0611 values between -1.06 and -0.9315 values between -0.93 and -0.8017 values between -0.80 and -0.6730 values between -0.67 and -0.5443 values between -0.54 and -0.4081 values between -0.40 and -0.27255 values between -0.27 and -0.141058 values between -0.14 and -0.011546 values between -0.01 and 0.12437 values between 0.12 and 0.25174 values between 0.25 and 0.3992 values between 0.39 and 0.5256 values between 0.52 and 0.6530 values between 0.65 and 0.7818 values between 0.78 and 0.9111 values between 0.91 and 1.044 values between 1.04 and 1.173 values between 1.17 and 1.301 values between 1.30 and 1.441 values between 2.22 and 2.36offnegativepositiveoften

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. etheless
  2. ilm
  3. 理
  4. ating
  5. 如今
  6. rical
  7. →⏎⏎
  8. .framework
  9. asc
  10. -)

Suppresses

  1. bole
  2. ncpy
  3. strap
  4. zaj
  5. Booker
  6. coat
  7. heit
  8. NTN
  9. :bold
  10. ncy