Neuron Explorer

Qwen3-0.6B

'a' after 'tell', 'bring' or 'much of'

A neuron in layer 5 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

The neuron fires on the word 'a' in phrases like 'tell a history', 'bring a fresh perspective' or 'much of a workout'. Turning it up did not give the writing a clear new subject.

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

    determine the movements necessary to pick up a weight in the invented environment. Grab hold, stiffen your arm, and lift. It doesn't sound like much of a workout, but the researchers say it could be useful in therapy for people recovering from

  2. #2 peak

    49F for pure metal and about 512F for tin-zinc alloys. Tests show that screws plated with the material do not stand much of a chance in hot areas. One test run several years ago by J. Bick

  3. #3 peak

    collections in Greater Manchester⏎A compass used in a PoW escape, a railway station sundial and part of the world's first commercial computer: objects that tell a history of Manchester and its place in the world.⏎BBC Manchester and museums across Greater

  4. #4 peak

    ofs is an introduction to modern symbolic logic, covering sentential and predicate logic (with identity). The course is highly interactive and engaging. It brings a fresh perspective to classical material by focusing on developing two crucial logical skills: strategic construction

  5. #5 peak

    Kitten? How Do I Tell a Kitten’s Age?⏎If your kitten is yet to get its adult teeth,

  6. #6 peak

    , must speak to a much wider audience: to the educated general public and academic audiences worldwide.⏎Inquiry editors are members of that wider audience, who can tell an author when a concept, process, or term is or is not clear. Often

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 question about problem of a 2D 2D problem of a problem of a 2D problem of a 2D problem of a 2D problem of a 2D problem of a 2D problem

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

    Moving Picture Experts Group⏎||Flexible, works with a wide range of editing and compression programs; have support for multiple video or audio streams⏎||MP4 is are usually compressed; compatible with more devices like iPad, iPod, iPhone, Android Phone

  2. #2 peak

    . R is a programming language for statistics computing. These tools, especially their open source packages, have

  3. #3 peak

    muscle tissue and white blood cells of autistic children, as well as in the brains of mice with a different mutation linked to autism2,3.⏎“We’ve had this global sense that mitochondrial dysfunction may be particularly relevant to synaptic defects, but I think

  4. #4 peak

    done with the bill. See HERE in the case of the Red-breasted Parakeet (Psittacula alexandri). On the other hand, with aboreal birds, they scratch with

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. L4:137 excites
  2. L4:2589 suppresses
  3. L3:1456 suppresses
  4. L4:52 excites
  5. L4:731 excites
  6. L3:2603 suppresses
  7. L4:2830 excites
  8. L4:2242 excites
  9. L4:283 weakly suppresses
  10. L3:597 weakly suppresses
  11. L0:1343 weakly excites
  12. L3:811 weakly excites
  13. L4:49 weakly excites
  14. L3:2843 weakly suppresses
  15. L0:173 weakly excites
  16. L0:1717 weakly suppresses
  17. L1:1684 weakly suppresses
  18. L2:699 weakly excites
  19. L3:1920 weakly excites
  20. L2:670 weakly excites
  21. L0:510 weakly excites
  22. L2:1446 weakly excites
  23. L0:2713 weakly suppresses
  24. L1:1078 weakly suppresses

Feeds later layers

  1. L16:2434 excites
  2. L8:1673 suppresses
  3. L23:728 suppresses
  4. L21:592 suppresses
  5. L17:188 suppresses 'a' before a noun phrase
  6. L6:530 excites
  7. L16:815 suppresses 'the' before a noun mid-sentence
  8. L8:466 suppresses 'a' or 'an' before a noun phrase
  9. L9:59 weakly suppresses "a" meaning per (a day, a year)
  10. L16:361 weakly suppresses
  11. L24:923 weakly excites
  12. L25:390 weakly suppresses
  13. L26:259 weakly suppresses
  14. L16:2701 weakly suppresses modifiers right before a noun
  15. L19:1060 weakly suppresses
  16. L23:2755 weakly excites
  17. L16:880 weakly suppresses
  18. L16:1309 weakly suppresses
  19. L25:837 weakly suppresses
  20. L8:595 weakly suppresses
  21. L26:110 weakly excites
  22. L20:901 weakly suppresses
  23. L14:990 weakly suppresses
  24. L19:1908 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 -2.92 and -2.781 values between -2.07 and -1.932 values between -1.50 and -1.361 values between -1.36 and -1.223 values between -1.08 and -0.943 values between -0.94 and -0.803 values between -0.80 and -0.665 values between -0.66 and -0.5210 values between -0.52 and -0.3752 values between -0.37 and -0.23543 values between -0.23 and -0.091942 values between -0.09 and 0.051070 values between 0.05 and 0.19194 values between 0.19 and 0.3337 values between 0.33 and 0.4710 values between 0.47 and 0.618 values between 0.61 and 0.767 values between 0.76 and 0.902 values between 0.90 and 1.046 values between 1.04 and 1.183 values between 1.18 and 1.321 values between 1.32 and 1.462 values between 1.46 and 1.601 values between 1.89 and 2.031 values between 2.59 and 2.73offnegativepositiveoften

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. ingham
  2. 等活动
  3. 的脚步
  4. edImage
  5. xff
  6. ingly
  7. 给出
  8. 别说
  9. illions
  10. 香味

Suppresses

  1. ignment
  2. aisal
  3. ⇑
  4. sweeping
  5. reement
  6. field
  7. UA
  8. 弧
  9. reass
  10. dynamic