Neuron Explorer

Qwen3-0.6B

first piece of a name starting 'Be'

A neuron in layer 24 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 fires on the start of a name, most often a surname, place or town that begins with the letters 'Be' or 'Op' or 'Ne', such as Beaufort, Bearden or Beagle. Turning it up did not produce any clear theme in 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

    on Coastal Hazards.⏎The detailed, 43-page document draws on tidal gauge measurements at five points along the eastern coast—at Duck, Oregon Inlet, Beaufort, Wilmington and Southport—as well as projections from the Intergovernment

  2. #2 peak

    specifically, the factors that may limit uptake–in order to shed light on how to improve the promotion of food literacy skills. We will also hear from Erin Beagle, Executive Director of Roots to Harvest, who will share her insights and experiences

  3. #3 peak

    wrote lives of Daniel Webster and James Buchanan.<|endoftext|>By: H Bernard Bechtel (Author)⏎206 pages, colour & b/w photos,

  4. #4 peak

    . (Image via Zach Montellaro on Flickr, CC BY-NC-SA 2.0)⏎The two Supreme Court cases, Jones v. City of Opelika 316 U.S. 584 (19

  5. #5 peak

    had this global sense that mitochondrial dysfunction may be particularly relevant to synaptic defects, but I think this [study] really provides more specificity to that story,” says Carrie Bearden, professor of psychiatry, biobehavioral sciences and psychology at

  6. #6 peak

    Origin of the Intentional Fallacy⏎The idea of the intentional fallacy was introduced by W.K. Wimsatt, Jr., and Monroe C. Beardsley in their book

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: ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' '

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

    about seven inches, tail a little more; and the Brown or Norway Rat (Mus decumanus) - the "Hanoverian "Rat of Waterton -

  2. #2 peak

    definition, we find that a pattern of distressing actions or demands<|endoftext|>Family Name: Passifloraceae⏎Botanical Name(s): Passiflora Incarnata⏎Popular Name(s): Maypops, Wild Passionflower and Apr

  3. #3 peak

    37th UGM doctoral graduate.⏎In conclusion of her dissertation entitled "Biosystematics of Murricata L. Annona, Annona squamosa and Annona Reticulata by Numerical Approach", it

  4. #4 peak

    a genus from the tropics, this plant has a kinship with Soursop (Annona muricata) and sugar apple (Annona squamosa). The custard-apple has become increasingly scarce because not many

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. L23:635 suppresses
  2. L21:1771 weakly suppresses
  3. L19:408 weakly excites
  4. L0:2723 weakly excites
  5. L17:2018 weakly suppresses
  6. L9:358 weakly suppresses
  7. L1:1662 weakly excites
  8. L4:2370 weakly suppresses
  9. L6:49 weakly suppresses
  10. L19:1409 weakly excites
  11. L15:222 weakly suppresses
  12. L23:293 weakly suppresses
  13. L19:1946 weakly suppresses
  14. L21:814 weakly suppresses
  15. L18:1879 weakly excites
  16. L14:1340 weakly excites
  17. L12:612 weakly excites
  18. L4:426 weakly excites
  19. L19:618 weakly excites
  20. L20:579 weakly excites
  21. L12:866 weakly suppresses
  22. L0:1935 weakly suppresses
  23. L22:1668 weakly excites
  24. L20:704 weakly excites

Feeds later layers

  1. L25:1641 suppresses
  2. L27:661 suppresses
  3. L27:315 suppresses
  4. L27:404 suppresses
  5. L27:1793 weakly excites
  6. L25:1922 weakly excites
  7. L25:2646 weakly excites 'there' after a because or if clause
  8. L25:2121 weakly excites
  9. L25:2166 weakly excites
  10. L25:776 weakly suppresses
  11. L25:1855 weakly suppresses
  12. L27:1786 weakly suppresses
  13. L25:2979 weakly suppresses
  14. L25:383 weakly suppresses
  15. L27:2932 weakly suppresses capital F starting a word
  16. L26:1175 weakly excites
  17. L25:1581 weakly excites
  18. L27:2731 weakly suppresses
  19. L26:2194 weakly excites
  20. L25:3068 weakly excites
  21. L27:1457 weakly excites
  22. L25:3013 weakly suppresses
  23. L27:618 weakly excites
  24. L26:302 weakly suppresses colon or comma before a quote

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.

2 values between -15.89 and -15.051 values between -15.05 and -14.201 values between -14.20 and -13.361 values between -11.67 and -10.823 values between -9.98 and -9.132 values between -9.13 and -8.291 values between -8.29 and -7.445 values between -6.60 and -5.753 values between -5.75 and -4.917 values between -4.91 and -4.063 values between -4.06 and -3.226 values between -3.22 and -2.378 values between -2.37 and -1.5357 values between -1.53 and -0.682966 values between -0.68 and 0.16783 values between 0.16 and 1.0131 values between 1.01 and 1.858 values between 1.85 and 2.702 values between 2.70 and 3.541 values between 3.54 and 4.394 values between 4.39 and 5.232 values between 5.23 and 6.081 values between 6.08 and 6.924 values between 6.92 and 7.772 values between 7.77 and 8.612 values between 8.61 and 9.461 values between 10.30 and 11.151 values between 17.06 and 17.91offnegativepositiveoften

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. ạn
  2. 孑
  3. нес
  4. 赘
  5. encion
  6. airl
  7. יתי
  8. <dd
  9. ->__
  10. обраща

Suppresses

  1. '
  2. "
  3. ‘
  4. “
  5. ('
  6. 「
  7. «
  8. \"
  9. ("
  10. '[