Neuron Explorer

Qwen3-0.6B

"to" after begin or start

A neuron in layer 4 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 the word "to" right after verbs like "begin," "began," or "start," as in "we begin to collect data." Its strongest negative cases are unrelated, such as line breaks in lecture transcripts. Turning it up did not produce a clear change in the model's replies.

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

    , and once we have a proof of concept that we can build a device that has these capabilities in principle, we can go to the testing stage where we begin to collect data from real epilepsy patients.⏎“Provid<|endoftext|>Dashes vs. Unders

  2. #2 peak

    puberty. “Many parents don’t want men to teach their daughters,” said Pushpa Kumar, the Okhla’s center’s teacher.<|endoftext|>When we begin to talk about "deferred" items in relation to accounting transactions, we begin to

  3. #3 peak

    by the Council of Chief State School Officers (CCSSO) and the National Governors Association (NGA).” My question is this: When did Montana begin to be represented by associations and councils rather than our elected representatives in the legislature? Denise

  4. #4 peak

    Performance measures for primary prevention of cardiovascular disease in adults: A report of the American College of Cardiology Foundation<|endoftext|>One of the interesting things that happens when one begins to coach others is that the skill being taught becomes embedded more deeply into one’s own

  5. #5 peak

    foresters talk about planting oaks throughout the state [and] having a forest products industry that begins to move toward oaks. If these come true, that's going to change the

  6. #6 peak

    is in those moments they become reenergize for and reinvested in further practice.⏎It is with this notion of practice in my head that I began to explore what practice is in education. How it is like practice in sport? I

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: if I can do this problem, and then I can use the knowledge that I have. It's a problem about how I can do this problem, and then I can use the knowledge that I have. It's a problem about how 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

    few common examples.…As I discussed in the video about signal flow,…there are three main categories…of the strength or level…of the electrical analog signal.…⏎Mic level, line level, and speaker level.…Mic level is the weakest of

  2. #2 peak

    .…It's less powerful than even a small…watch battery.…For processing and mixing,…the weak mic level signal…needs a microphone preamplifier,…or mic preamp, to boost it

  3. #3 peak

    is useful when we do our u-substitution,⏎- we could say that du is equal to 5 cosine 5x.⏎- Now when you look over here, we don't have quite du there. We have just cosine of 5

  4. #4 peak

    Candles are made from soybeans, a natural, renewable source grown in many regions by local farmers....right here in the USA! A candle made from soy wax burns cleaner, lasts longer and

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. L3:191 suppresses
  2. L3:213 suppresses
  3. L3:59 excites
  4. L2:493 suppresses
  5. L1:2830 weakly suppresses
  6. L1:2459 weakly excites
  7. L3:456 weakly suppresses
  8. L1:939 weakly suppresses
  9. L3:179 weakly excites
  10. L2:111 weakly suppresses
  11. L1:2658 weakly suppresses
  12. L1:1464 weakly excites
  13. L0:178 weakly suppresses
  14. L3:1311 weakly suppresses
  15. L1:74 weakly excites
  16. L3:2369 weakly suppresses
  17. L2:70 weakly excites
  18. L0:1224 weakly suppresses
  19. L2:2116 weakly excites
  20. L2:1036 weakly excites
  21. L2:1599 weakly excites
  22. L0:1301 weakly suppresses
  23. L3:50 weakly excites
  24. L1:23 weakly suppresses

Feeds later layers

  1. L9:776 suppresses
  2. L9:2779 suppresses
  3. L6:49 excites
  4. L5:1047 suppresses
  5. L20:2621 suppresses
  6. L9:1626 suppresses
  7. L23:65 excites
  8. L8:874 weakly suppresses
  9. L16:123 weakly excites
  10. L18:1766 weakly suppresses
  11. L9:1149 weakly suppresses
  12. L10:1464 weakly excites
  13. L5:2325 weakly suppresses
  14. L16:2405 weakly suppresses
  15. L12:647 weakly excites
  16. L12:2218 weakly suppresses
  17. L7:1700 weakly suppresses
  18. L7:1979 weakly excites
  19. L16:89 weakly excites
  20. L9:1533 weakly suppresses
  21. L9:627 weakly excites
  22. L22:5 weakly excites
  23. L7:1843 weakly excites
  24. L15:73 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 -0.86 and -0.761 values between -0.76 and -0.661 values between -0.66 and -0.574 values between -0.57 and -0.472 values between -0.47 and -0.3816 values between -0.38 and -0.2881 values between -0.28 and -0.19370 values between -0.19 and -0.091188 values between -0.09 and 0.011271 values between 0.01 and 0.10590 values between 0.10 and 0.20250 values between 0.20 and 0.2992 values between 0.29 and 0.3929 values between 0.39 and 0.497 values between 0.49 and 0.583 values between 0.58 and 0.681 values between 0.87 and 0.961 values between 2.88 and 2.98offnegativepositiveoften

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. prints
  2. 锰
  3. stands
  4. -fin
  5. 受限
  6. TEMPL
  7. kdir
  8. leaf
  9. 름
  10. uced

Suppresses

  1. IPA
  2. osemite
  3. ALA
  4. Vimeo
  5. ynchronize
  6. quelle
  7. gli
  8. translating
  9. .translate
  10. iota