Neuron Explorer

Qwen3-0.6B

comma after a clause subject

A neuron in layer 2 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 a comma that comes right after the subject of a sentence, just before the rest of the sentence continues, such as 'When a recycled material, rather than...'. When it was turned up, the model's text did not visibly change in a clear way.

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

    adeption.⏎After Edward IV’s death, Elizabeth tried to mediate between the factions on the king’s council. At the news that her brother and her son, Richard Grey, had been arrested by Richard Duke of Gloucester, she at first

  2. #2 peak

    little isle of Oronsay, close by Islay; and here for a moment they thought their future abode was to be. But when Columba, with the early morning, ascending the highest ground on the island, to take what

  3. #3 peak

    economic benefit?⏎A: In a broad sense, recycling is part of an ethic of resource efficiency – of using products to their fullest potential. When a recycled material, rather than a raw<|endoftext|>Toss it. Pitch it. Dump it. Chuck

  4. #4 peak

    0, and it won’t arrive at the Jezero Crater until February 18, 2021. But if the helicopter soars, it’ll provide never-before-seen views of the Red Planet.⏎Popayán

  5. #5 peak

    rich blend of cultural and traditional customs. All Americans are renowned for their openness and friendliness to visitors. As a rule, dress is casual; although smart restaurants, hotels and clubs insist on suits and ties or long dresses. Smoking is becoming increasingly

  6. #6 peak

    academic disciplines that contribute to HCI research. The second part of the book focuses on the application of HCI to Information Systems research, and reviews ways in which HCI techniques, methodologies and other research components have been used to date in the IS field. The

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: "I have a lot of time, and the last I have been thinking about lately. This is a fun way to learn about, and the last I have been thinking about lately. I have been thinking about the times and the last I

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

    focuses on all delinquent youth in the community, while the second examines only delinquent youth placed in juvenile pretrial detention centers when

  2. #2 peak

    Trail and ‘The Chief’ viewing platforms.⏎Take a virtual tour of the Sea-to-Sky viewing platform…⏎To find out more about the Britannia Mining Museum please visit: http://www.britanniaminemuseum.ca/⏎To find out

  3. #3 peak

    to create the next-generation processor, while maintaining backward compatibility and ease of replication; that is, Intel had to be able

  4. #4 peak

    ⏎MRI scans can detect very small irregularities and produce a very high quality picture on its own, while the operator may also give the patient a contrast agent that can make the image even

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. L1:546 excites
  2. L1:171 suppresses "but" after a negated clause
  3. L1:383 weakly suppresses
  4. L1:105 weakly excites
  5. L1:505 weakly excites
  6. L1:1231 weakly suppresses
  7. L1:89 weakly excites
  8. L1:211 weakly excites
  9. L0:27 weakly suppresses
  10. L1:125 weakly excites
  11. L1:644 weakly suppresses
  12. L1:807 weakly excites
  13. L0:2877 weakly suppresses
  14. L1:698 weakly suppresses
  15. L1:1149 weakly suppresses
  16. L1:453 weakly excites
  17. L1:603 weakly excites
  18. L1:694 weakly excites
  19. L0:2582 weakly suppresses
  20. L1:1344 weakly suppresses 'who' after a person noun
  21. L1:1001 weakly excites
  22. L1:1715 weakly suppresses
  23. L1:2677 weakly suppresses
  24. L0:585 weakly suppresses

Feeds later layers

  1. L3:182 excites
  2. L4:80 suppresses
  3. L3:227 suppresses
  4. L3:524 excites
  5. L4:44 suppresses
  6. L3:47 suppresses
  7. L3:569 excites unclear: mid-phrase function words
  8. L5:420 suppresses
  9. L6:352 suppresses
  10. L4:69 weakly suppresses
  11. L4:317 weakly excites
  12. L3:518 weakly excites
  13. L4:59 weakly suppresses
  14. L24:2951 weakly excites
  15. L4:305 weakly suppresses
  16. L22:1702 weakly excites
  17. L5:1598 weakly suppresses
  18. L15:157 weakly suppresses
  19. L11:1285 weakly suppresses
  20. L17:27 weakly suppresses
  21. L11:2917 weakly excites
  22. L8:2210 weakly suppresses
  23. L4:539 weakly suppresses
  24. L20:82 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 -1.66 and -1.591 values between -1.37 and -1.301 values between -1.30 and -1.221 values between -1.22 and -1.151 values between -1.15 and -1.082 values between -1.08 and -1.004 values between -0.93 and -0.863 values between -0.86 and -0.783 values between -0.78 and -0.7113 values between -0.71 and -0.639 values between -0.63 and -0.5623 values between -0.56 and -0.4922 values between -0.49 and -0.4131 values between -0.41 and -0.3455 values between -0.34 and -0.2796 values between -0.27 and -0.19231 values between -0.19 and -0.12523 values between -0.12 and -0.051161 values between -0.05 and 0.03811 values between 0.03 and 0.10462 values between 0.10 and 0.17272 values between 0.17 and 0.25100 values between 0.25 and 0.3234 values between 0.32 and 0.3915 values between 0.39 and 0.476 values between 0.47 and 0.546 values between 0.54 and 0.619 values between 0.61 and 0.695 values between 0.69 and 0.761 values between 0.83 and 0.911 values between 0.91 and 0.983 values between 0.98 and 1.051 values between 1.05 and 1.131 values between 1.20 and 1.27offnegativepositiveoften

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. 量
  2. ",
  3. imony
  4. tin
  5. )":
  6. ibrate
  7. TRA
  8. ivities
  9. いで
  10. طا

Suppresses

  1. MPL
  2. ukkit
  3. jp
  4. _PHP
  5. 길
  6. /slick
  7. =path
  8. ilenames
  9. ingular
  10. ActiveSupport