Neuron Explorer

Qwen3-0.6B

comma before a 'because' or 'as' clause

A neuron in layer 25 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 lights up on a comma that comes right before a reason, such as 'not a substitute, as this will be toxic'. When it was turned up, the model kept writing the word 'as' over and over.

Written by Claude from this neuron's activations and steering outputs. Confidence: high. 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

    this happens, add a watering agent such as Water-In to enhance soil penetration. Water-In is available at most garden stores. Do not use detergent as a substitute, as this will be toxic to the plant.⏎Heavy clay soils drain slowly. If

  2. #2 peak

    organised, and properly targeted for temperature-controlled shipping and storage.⏎First, the purpose of the kit or kits must be firmly defined. This is a critical stage, as imprecision in

  3. #3 peak

    data, only 2 out of 10 households meet the estimated average requirement (EAR) of Vitamin A in their daily diet. This trend is particularly worrying, given the rising rates of food insecurity that arose as a consequence of the COVID-

  4. #4 peak

    to wipe the fridge.⏎· Avoid using any soap or detergent, as they might leave behind a residue or scent, which can contaminate your

  5. #5 peak

    is made up of swirling droplets of water. Many people assume that a waterspout is nothing more than a tornado over water. This is not entirely true, however, as some waterspouts can develop in relatively fair weather conditions unlike the

  6. #6 peak

    ores⏎This is a fun activity to do at the beginning (or end) of the school year. Don’t attempt this on a cool or overcast day, because you need heat and direct sunlight to cook the s’mores. You will

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 the pastes and pastes are because of the pastes because of the pastes... or the pastes because of the pastes and the pastes because of the pastes. I'm not sure if this is the right way

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

    U.C. Mohanty, a professor in the Centre for Atmospheric Sciences at the Indian Institute of Technology, and Mohanty's doctoral student, Ashish Routray, collaborated with scientists at the National Center for Atmospheric Research and gathered information

  2. #2 peak

    examples of air pollution.<|endoftext|>Revealing the art and architecture of Vijayanagara Empire, Vittala Temple Complex is one of the major tourist attractions in Hampi. The complex and its structures embody the pinnacle of Dravidian temple architectural

  3. #3 peak

    D) ministry. Rather, the demand follows from a chest thumping and energetic flag waving ‘Tiranga March’ that was recently held at the Jawaharlal Nehru University (JNU).⏎The

  4. #4 peak

    . During these ten years he emerged as the overseas leader of the Indonesian nationalist movement and became acquainted with counterparts representing other independence movements, including the Indian leader Jawaharlal Nehru. In 1927 Hatta was accused by Dutch

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. L22:2261 suppresses
  2. L22:1385 suppresses
  3. L22:592 excites
  4. L21:2115 excites
  5. L21:1111 suppresses
  6. L24:951 excites
  7. L22:2677 excites
  8. L22:1883 excites
  9. L22:684 suppresses
  10. L19:710 suppresses
  11. L24:1392 suppresses
  12. L19:1570 excites
  13. L10:39 excites
  14. L18:1444 suppresses
  15. L20:541 suppresses
  16. L5:136 excites
  17. L19:2495 excites
  18. L20:1043 excites
  19. L21:2669 suppresses
  20. L23:1635 excites
  21. L19:1036 suppresses
  22. L23:986 excites
  23. L10:190 weakly excites
  24. L21:2563 weakly suppresses

Feeds later layers

  1. L26:588 strongly excites
  2. L27:471 strongly excites
  3. L26:2423 excites
  4. L26:790 suppresses
  5. L27:811 excites
  6. L26:1024 weakly excites
  7. L26:2115 weakly suppresses
  8. L26:2926 weakly suppresses
  9. L26:1320 weakly suppresses
  10. L27:1497 weakly suppresses
  11. L26:2152 weakly suppresses
  12. L26:220 weakly suppresses
  13. L27:1676 weakly suppresses
  14. L27:493 weakly suppresses
  15. L27:1227 weakly suppresses
  16. L27:1082 weakly suppresses
  17. L26:953 weakly suppresses
  18. L27:1406 weakly suppresses
  19. L26:2742 weakly suppresses
  20. L26:2123 weakly suppresses
  21. L26:218 weakly excites
  22. L26:2286 weakly excites
  23. L27:1588 weakly excites
  24. L27:2961 weakly excites

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 -9.69 and -8.841 values between -7.99 and -7.141 values between -7.14 and -6.281 values between -6.28 and -5.434 values between -5.43 and -4.584 values between -4.58 and -3.731 values between -3.73 and -2.889 values between -2.88 and -2.0317 values between -2.03 and -1.18223 values between -1.18 and -0.333105 values between -0.33 and 0.52484 values between 0.52 and 1.3729 values between 1.37 and 2.2210 values between 2.22 and 3.077 values between 3.07 and 3.932 values between 3.93 and 4.781 values between 4.78 and 5.631 values between 5.63 and 6.481 values between 6.48 and 7.331 values between 8.18 and 9.031 values between 9.03 and 9.881 values between 9.88 and 10.731 values between 11.58 and 12.431 values between 20.09 and 20.941 values between 23.49 and 24.34offnegativepositiveoften

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. as
  2. 作为
  3. →as
  4. As
  5. 作為
  6. As
  7. -as
  8. 作为一个
  9. as
  10. _as

Suppresses

  1. 因为在
  2. ragments
  3. because
  4. 因为你
  5. ーン
  6. 因为他
  7. 因为
  8. because
  9. pois
  10. anchise