Neuron Explorer

Qwen3-0.6B

preposition 'in' after a noun

A neuron in layer 20 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 mostly on the word 'in' when it follows a noun or verb, as in 'involved in' or 'in our everyday lives'. Turned up, the model got stuck repeating the word 'that', which is a different thing it pushes toward.

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

    the comfort of your own home, at your computer. With identity theft posing a real threat, though, keeping financial data private requires that consumers be proactive in the way that they approach online security.⏎The Rising Wave of Identity Theft⏎According to the

  2. #2 peak

    many students, these choices remain a reality for survival at the Institute. MIT students too often sacrifice sleep, exercise, social activity, and other forms of personal care in order to enhance their educational experience. The consequences of this lifestyle imbalance can be far

  3. #3 peak

    000 words— past and present—from across the English-speaking world. As a historical dictionary, the OED is very different from those of current English, in which the focus is on present-day meanings. You’ll still find these in the

  4. #4 peak

    with the FDA. It must continue all the way to our dinner tables, which means that we, as consumers, need to know and practice proper food safety habits in our everyday lives. Here are

  5. #5 peak

    of animal welfare advocates.”<|endoftext|>People that live in islands still use water transport as a means of transportation. People that rely on fishing for their livelihood use water transport in fishing. A boat is widely used in sea transport. A boat is more flexible

  6. #6 peak

    “any breeder, public or private, the opportunity to send germplasm they wish to have evaluated. The information we generate is useful to the breeders in knowing the strengths and weaknesses of potential cultivars. The data is also critical in

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: to use a todi to make a todi to do a todi to do ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...

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

    treat each other.⏎Intervene Quickly & Briefly⏎Many adults tell me that when they witness cruelty in their classrooms, they freeze up and don’t know what to do or say. I tell them that the most effective way to intervene is

  2. #2 peak

    9)⏎practice, or doing. Reality is manifested in concrete situations.⏎coming to terms with things by studying words individually.⏎choosing what perspective to take in a situation; the unveiling of an essential meaning.⏎approaches in

  3. #3 peak

    verification code via phone call⏎- SMS – the user can receive a text message containing the verification code or password reset link⏎When it comes time to decide which method(s) to use for

  4. #4 peak

    through all of it.⏎So what’s the key? Goals have to be actionable. That is the only way to make sure they will know what steps to take, and don’t waste valuable time spinning their wheels.⏎Does your teen

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. L18:2666 suppresses
  2. L17:2582 excites
  3. L18:2740 suppresses
  4. L19:1235 suppresses
  5. L19:2621 suppresses
  6. L18:1389 excites
  7. L18:2031 suppresses
  8. L18:1617 weakly excites
  9. L17:1282 weakly excites
  10. L17:2783 weakly suppresses
  11. L15:2692 weakly excites 'to' after a noun, before a purpose verb
  12. L14:2528 weakly suppresses
  13. L3:50 weakly excites
  14. L19:2737 weakly excites
  15. L13:164 weakly excites
  16. L7:1152 weakly suppresses
  17. L17:1368 weakly suppresses
  18. L19:945 weakly suppresses
  19. L17:2695 weakly suppresses
  20. L19:3040 weakly excites
  21. L9:1872 weakly suppresses
  22. L17:1884 weakly suppresses
  23. L18:2575 weakly excites
  24. L7:1024 weakly excites

Feeds later layers

  1. L22:1502 suppresses
  2. L24:959 suppresses
  3. L21:1550 excites
  4. L23:830 excites
  5. L27:616 excites
  6. L22:892 excites
  7. L26:586 excites
  8. L22:631 excites
  9. L22:2014 excites
  10. L26:408 suppresses
  11. L22:1186 excites
  12. L21:2876 weakly excites
  13. L23:382 weakly suppresses
  14. L21:2711 weakly suppresses
  15. L24:381 weakly suppresses
  16. L23:914 weakly excites
  17. L23:2514 weakly excites
  18. L21:1422 weakly excites
  19. L22:1616 weakly excites
  20. L21:639 weakly excites
  21. L23:298 weakly excites
  22. L21:1851 weakly excites
  23. L22:181 weakly excites
  24. L21:1521 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.

2 values between -9.36 and -8.762 values between -8.16 and -7.551 values between -7.55 and -6.951 values between -6.95 and -6.352 values between -6.35 and -5.754 values between -5.75 and -5.142 values between -5.14 and -4.545 values between -4.54 and -3.942 values between -3.94 and -3.343 values between -3.34 and -2.748 values between -2.74 and -2.1311 values between -2.13 and -1.5334 values between -1.53 and -0.93760 values between -0.93 and -0.332611 values between -0.33 and 0.28259 values between 0.28 and 0.8877 values between 0.88 and 1.4833 values between 1.48 and 2.0819 values between 2.08 and 2.6820 values between 2.68 and 3.2911 values between 3.29 and 3.898 values between 3.89 and 4.495 values between 4.49 and 5.094 values between 5.09 and 5.697 values between 5.69 and 6.302 values between 6.30 and 6.901 values between 6.90 and 7.503 values between 7.50 and 8.103 values between 8.71 and 9.311 values between 9.31 and 9.911 values between 9.91 and 10.511 values between 10.51 and 11.112 values between 11.11 and 11.721 values between 11.72 and 12.321 values between 12.32 and 12.921 values between 14.12 and 14.73offnegativepositiveoften

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. that
  2. that
  3. dass
  4. →that
  5. rằng
  6. że
  7. _that
  8. sofar
  9. bahwa
  10. (that

Suppresses

  1. ki
  2. udio
  3. udios
  4. to
  5. quets
  6. .to
  7. ilion
  8. 提出了
  9. icious
  10. riet