Neuron Explorer

Qwen3-0.6B

capital E starting 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 responds to a capital E that opens an acronym in brackets, like the E in '(EIGA)' or '(ECD)'. When it is pushed negative it tends to respond to a lowercase e starting a word fragment instead. Steering outputs were not clearly connected to this 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

    tissue.⏎- Pull the<|endoftext|>Listen to radar echoes from satellites and meteors, live on listener-supported Space Weather Radio.⏎| || |⏎ORIONID METEOR UPDATE: Earth is passing through a broad stream of debris from Halley's

  2. #2 peak

    1; G-1.2; G-1.6; G-1.9; G1.10, European Industrial Gases Association (EIGA) – IGC Doc 123/1<|endoftext|>How Canadian

  3. #3 peak

    researchers and regular people can use to do science and educate.<|endoftext|>The learning program at Rona Glynn Preschool is designed using the Early Years Learning Framework (EYLFv2.0) Australia’s national Early Years Framework used cohes

  4. #4 peak

    combined with Scavenge Air Moisturising (SAM) or Exhaust Gas Recirculation (EGR) systems are evaluated, as such systems will probably be applied on future engines

  5. #5 peak

    to assist you during this process.⏎State and District Resources⏎Comprehensive School Reform: Five Lessons from the Field -- Available from the Education Commission of the States (ECS) Education

  6. #6 peak

    food was haraam, his drink was haraam, his clothing was haraam,<|endoftext|>You may have passed your International English Language Testing System (EILTS), TOEFL, TOEIC, a college admission essay

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: When people keeppickingpickingpickingpickingpickingpickingpickingpickingpickingpickingpickingpickingpickingpickingpickingpickingpickingpickingpickingpickingpickingpickingp

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

    patient care," Venneti says. The study is published in Science Translational Medicine.⏎Traditional prognostic markers based on tumor grade have proven unreliable for ependymomas. Understanding a patient's prognosis can help doctors make treatment recommendations,

  2. #2 peak

    ) is the UK government agency that helps provide this capability.⏎Satellite data is essential for comparing to computer simulation models, historical (paleo) data from e.g. ice cores and tree rings, and to test and improve our scientific understanding

  3. #3 peak

    has been associated largely with human intervention on coastal processes in the form of industrialization, recreation, watershed runoff, agricultural land uses, dredge and fill operations, eutrophication, and unsustainable fishing practices;⏎Recalling the catastrophic consequences to fisheries

  4. #4 peak

    terrestrial discharge, atmospheric inputs, mixing, and ecological interactions. As that process unfolds, interim models representing the best understanding to date are applied to real world problems of eutrophication and water quality created by human habitation of the

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. L18:221 excites
  2. L22:2428 excites
  3. L17:542 suppresses
  4. L16:56 suppresses
  5. L21:361 suppresses
  6. L17:594 suppresses
  7. L2:1749 excites
  8. L19:598 excites
  9. L20:374 suppresses
  10. L24:586 excites
  11. L20:1398 excites
  12. L0:2901 excites
  13. L10:2401 excites
  14. L16:118 suppresses
  15. L0:54 excites
  16. L17:366 suppresses
  17. L14:392 weakly suppresses
  18. L16:205 weakly excites
  19. L1:141 weakly suppresses
  20. L18:1104 weakly excites
  21. L19:645 weakly suppresses
  22. L2:1795 weakly excites
  23. L18:139 weakly excites
  24. L0:1242 weakly suppresses

Feeds later layers

  1. L27:2255 weakly suppresses
  2. L27:2880 weakly suppresses
  3. L27:736 weakly suppresses
  4. L27:2763 weakly excites
  5. L27:1436 weakly suppresses
  6. L27:2963 weakly suppresses the 'st' piece inside words
  7. L27:758 weakly suppresses
  8. L27:1427 weakly suppresses
  9. L27:1810 weakly suppresses
  10. L27:1307 weakly excites
  11. L27:1849 weakly suppresses
  12. L27:2188 weakly suppresses
  13. L27:1147 weakly suppresses
  14. L27:1469 weakly excites
  15. L27:1802 weakly suppresses
  16. L27:2389 weakly suppresses
  17. L27:2597 weakly suppresses
  18. L27:2417 weakly excites
  19. L27:2452 weakly excites
  20. L27:1773 weakly suppresses
  21. L27:2909 weakly excites
  22. L27:2662 weakly suppresses
  23. L27:1476 weakly suppresses
  24. L27:792 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 -84.56 and -81.041 values between -24.62 and -21.101 values between -21.10 and -17.573 values between -17.57 and -14.053 values between -14.05 and -10.524 values between -10.52 and -7.0014 values between -7.00 and -3.472277 values between -3.47 and 0.061585 values between 0.06 and 3.589 values between 3.58 and 7.113 values between 7.11 and 10.632 values between 10.63 and 14.163 values between 14.16 and 17.691 values between 17.69 and 21.211 values between 52.94 and 56.47offnegativepositiveoften

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. 就好像
  2. ièrement
  3. ['',
  4. "'",
  5. .pem
  6. ục
  7. i
  8. 涪
  9. .fd
  10. unte

Suppresses

  1. mö
  2. 翰
  3. 全国人民
  4. mer
  5. finite
  6. p
  7. m
  8. nable
  9. motion
  10. Gen