Neuron Explorer

Qwen3-0.6B

unclear: word before a twist or contrast

A neuron in layer 17 of Qwen3-0.6B. Its strongest snippets are mixed: they peak on different tokens. It is featured, so it has steering outputs.

What this neuron does

This neuron fires on a word in the middle of a sentence where the writing seems to be setting up a turn, like 'far from over', 'doesn't end there' or 'wasn't the first'. It is hard to pin down more precisely. Turning it up made the model's continuations wander off into unrelated topics.

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

    technology status, the fuel cell has experienced an astounding development. Fuel cells are already being used in a variety of applications today. But its impressive career is far from over. Because of their simple operation, the use of fuel cells in electric vehicles represents

  2. #2 peak

    Dell Encrypt?⏎Dell encrypts all personal information you give us when you place an order online. Remember, if you feel uncomfortable providing any of this information online, please feel free to call one of our representatives at 1-800

  3. #3 peak

    a 10-fold increase in storage capacity as well as a 300x improvement in performance over today’s drives.⏎The good news doesn’t end there, though. Even though this new technique relies on heat, it is thought to

  4. #4 peak

    year of learning, students often spend their summers not thinking about schoolwork and lose a good amount of the learning progress they have made. At Schoolmart, we understand the importance of continuous learning and are determined to bridge the summer knowledge gap with helpful

  5. #5 peak

    famous Paul who is credited with writing much of the New Testament.)⏎Although all the original 12 apostles are important, some get top billing. Here’s why:⏎

  6. #6 peak

    were less hardy than their parents. The same can be said about Dolly. She aged prematurely and passed away after 6 years.⏎Dolly wasn’t the first vertebrate to be cloned. That honour belongs to a carp, cloned 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: the first time, you have to be careful with the first time you have to be careful with the first time you have to be careful with the first time you have to be careful with the first time you have to be careful with the first

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

    have to try slow your heart down so you can hit a target that is about the size of a coin. I don't know how they do it. It sounds really difficult.⏎Racers wear cross-country skis, ski

  2. #2 peak

    to read. Reading is fascinating. It's full of surprises. And it will take you places you can't get to in any other way.⏎- Do EACH of the following:⏎- Learn how

  3. #3 peak

    can be as simple as a long alkyl chain, or a short stretch of ethylene glycol ethers, up to as complex a linker as you have the patience to search for.⏎It seems like a rather simple idea that you could get two

  4. #4 peak

    of Frank Churchill’s secret engagement to Jane.⏎“Good God!” cried Emma, “Jane actually on the point of going as governess! What could he mean by such horrible indelicacy? To suffer her to engage herself

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. L15:484 excites
  2. L15:348 suppresses
  3. L10:2688 suppresses
  4. L10:1610 excites
  5. L13:2710 suppresses
  6. L10:976 suppresses
  7. L14:1580 excites
  8. L10:375 excites
  9. L15:127 excites
  10. L16:1308 excites
  11. L10:380 excites
  12. L12:2444 excites
  13. L15:1803 excites
  14. L5:374 excites
  15. L10:1320 suppresses
  16. L9:1234 excites
  17. L11:772 excites
  18. L12:1831 weakly excites
  19. L14:3041 weakly excites
  20. L10:1289 weakly excites
  21. L14:434 weakly excites
  22. L11:961 weakly excites
  23. L16:632 weakly suppresses
  24. L8:2370 weakly suppresses

Feeds later layers

  1. L19:1653 excites
  2. L21:1972 weakly excites
  3. L18:2050 weakly excites
  4. L18:2722 weakly excites
  5. L18:2636 weakly excites
  6. L18:2704 weakly excites
  7. L22:2880 weakly excites
  8. L23:1227 weakly excites
  9. L19:1620 weakly excites
  10. L25:2960 weakly excites
  11. L22:705 weakly excites
  12. L20:2470 weakly suppresses
  13. L21:2358 weakly excites
  14. L19:585 weakly suppresses
  15. L19:3047 weakly excites
  16. L19:1868 weakly excites
  17. L21:1945 weakly excites
  18. L18:2545 weakly excites
  19. L20:2265 weakly suppresses
  20. L21:679 weakly excites
  21. L22:1390 weakly excites
  22. L19:2552 weakly excites
  23. L20:802 weakly excites
  24. L23:2716 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 -7.38 and -7.021 values between -6.65 and -6.281 values between -6.28 and -5.911 values between -5.91 and -5.542 values between -5.54 and -5.171 values between -5.17 and -4.815 values between -4.81 and -4.444 values between -4.44 and -4.076 values between -4.07 and -3.7010 values between -3.70 and -3.3317 values between -3.33 and -2.9723 values between -2.97 and -2.6023 values between -2.60 and -2.2332 values between -2.23 and -1.8656 values between -1.86 and -1.4944 values between -1.49 and -1.1278 values between -1.12 and -0.76129 values between -0.76 and -0.39271 values between -0.39 and -0.02664 values between -0.02 and 0.35656 values between 0.35 and 0.72546 values between 0.72 and 1.09453 values between 1.09 and 1.45349 values between 1.45 and 1.82215 values between 1.82 and 2.19122 values between 2.19 and 2.5673 values between 2.56 and 2.9358 values between 2.93 and 3.2931 values between 3.29 and 3.669 values between 3.66 and 4.0313 values between 4.03 and 4.406 values between 4.40 and 4.771 values between 4.77 and 5.143 values between 5.14 and 5.502 values between 5.50 and 5.871 values between 5.87 and 6.241 values between 6.98 and 7.34offnegativepositiveoften

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. ois
  2. '`
  3. iores
  4. addock
  5. UDIO
  6. aised
  7. ValidateAntiForgeryToken
  8. reetings
  9. てくれ
  10. '!

Suppresses

  1. 下午
  2. бан
  3. *ft
  4. erg
  5. .cy
  6. ification
  7. -ph
  8. bus
  9. scheduler
  10. 通过