Neuron Explorer

Qwen3-0.6B

opening parenthesis before an acronym

A neuron in layer 26 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 an opening bracket that comes right after a full name and introduces its short form, like 'Finite Element Method (FEM)'. Turned up, it made the model repeat words like 'aur', which does not clearly match that pattern.

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

    , and living or life skills training.⏎- Eligible California residents may connect with the San Andreas Regional Center (SARC) to create an Individualized Program Plan (IPP). This is a contract between SARC and the individual that sets up an

  2. #2 peak

    from injuring themselves and others by requiring gun owners to keep guns locked up and unloaded, and ammunition stored in a locked location separate from the firearm.⏎Child Access Prevention (CAP) laws go one step further by imposing criminal liability to ensure that owners are

  3. #3 peak

    companies that includes Advanced Plasma Power (APP), is building a full-scale gas plant that will prove this technology is commercially

  4. #4 peak

    have their influence on one another visually mapped. This methodology was named Interpretive Structural Modelling (ISM). Over the next 20 years, the use of ISM to

  5. #5 peak

    Technologies.⏎Dr Copeland and co-investigator Dr Carl Sangan will assess the technical feasibility of producing high temperature radial turbines using the Selective Laser Melting (SLM) Additive Manufacturing (AM) technique with nickel super-alloys.⏎

  6. #6 peak

    , we enhance these assumptions by considering mechanically diverse layers with varying dips. We employed the Finite Element Method (FEM) to create numerical models simulating two chamber shapes: a circular form

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: there's one data sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera sera

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

    Heating, Refrigeration and Air-Conditioning Engineers (ASHRAE)⏎Structural Engineers Association of Kansas & Missouri (SEAKM)⏎Structural Engineering Institute (SEI)⏎Illuminating Engineering Society (IES)⏎American Society of Plumbing Engineers

  2. #2 peak

    Propulsion Laboratory manages the 2001 Mars Odyssey mission for NASA's Office of Space Science, Washington, D.C. The Thermal Emission Imaging System (THEMIS) was developed by Arizona State University, Tempe, in collaboration

  3. #3 peak

    Council (ESRC), the research was presented recently at a day-long higher education stakeholder event in Johannesburg, hosted by the team and the Council for Higher Education (CHE). The MOOC was launched at the same event, which was the first

  4. #4 peak

    mental illness for good.<|endoftext|>Researchers synthesized the literature on physical security planning from school-specific sources and related disciplines to discuss the policy and other challenges that local education agencies (LEAs) face in developing and implementing protection and mitigation measures for kindergarten-through–

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. L25:2055 suppresses
  2. L24:600 excites
  3. L25:925 excites
  4. L23:625 suppresses
  5. L21:468 excites
  6. L20:1423 suppresses
  7. L19:1422 excites
  8. L22:309 excites
  9. L23:176 suppresses
  10. L24:120 suppresses
  11. L20:2497 suppresses
  12. L21:2936 suppresses
  13. L22:1393 suppresses
  14. L23:666 excites
  15. L22:982 excites
  16. L20:2479 suppresses
  17. L24:2046 excites opening bracket before an acronym
  18. L24:241 suppresses
  19. L21:1400 suppresses
  20. L19:2585 suppresses
  21. L12:642 excites
  22. L25:586 suppresses
  23. L19:1317 excites
  24. L25:2082 suppresses first-person 'I' starting a clause

Feeds later layers

  1. L27:563 strongly suppresses
  2. L27:1781 suppresses
  3. L27:2008 suppresses
  4. L27:273 weakly suppresses
  5. L27:2740 weakly excites
  6. L27:897 weakly suppresses
  7. L27:2870 weakly excites
  8. L27:913 weakly excites
  9. L27:536 weakly suppresses
  10. L27:760 weakly excites
  11. L27:630 weakly excites
  12. L27:2929 weakly excites the letter c splitting a word
  13. L27:915 weakly suppresses unclear: mixed word fragments and phrases
  14. L27:957 weakly excites
  15. L27:1009 weakly excites opening parenthesis before an acronym
  16. L27:2324 weakly excites
  17. L27:1942 weakly excites
  18. L27:1200 weakly excites
  19. L27:3033 weakly suppresses
  20. L27:1283 weakly excites
  21. L27:3038 weakly excites
  22. L27:516 weakly excites
  23. L27:2821 weakly excites
  24. L27:709 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 -13.12 and -12.071 values between -9.97 and -8.922 values between -6.82 and -5.774 values between -4.73 and -3.683 values between -3.68 and -2.633 values between -2.63 and -1.5871 values between -1.58 and -0.533741 values between -0.53 and 0.5267 values between 0.52 and 1.574 values between 1.57 and 2.623 values between 2.62 and 3.672 values between 4.72 and 5.771 values between 5.77 and 6.811 values between 7.86 and 8.911 values between 8.91 and 9.962 values between 11.01 and 12.061 values between 27.80 and 28.84offnegativepositiveoften

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. arrêt
  2. aur
  3. 格尔
  4. clid
  5. Athen
  6. jem
  7. eton
  8. vrir
  9. utes
  10. öt

Suppresses

  1. sera
  2. DA
  3. da
  4. ва
  5. HA
  6. aura
  7. ha
  8. VA
  9. inda
  10. leta