Neuron Explorer

Qwen3-0.6B

space between month and day

A neuron in layer 6 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 on the blank space between a month name and the day number, as in 'April 22' or 'July 20 1969'. Its opposite reaction is on the last digit of a year, like the '9' in '1969'. Turning it up did not produce a clear theme, so steering doesn't confirm the pattern. Its activations are fairly small compared with the other neurons here.

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

    similar study model<|endoftext|>Celebrate Earth Day 2012Certification Map, Education, General Interest | No Comments »⏎Earth Day is celebrated on April 22, and this year marks the 42nd anniversary

  2. #2 peak

    1863. In the two World Wars, Fort Hamilton served as a major embarkation and separation center. Most recently, in the aftermath of the September 11th terrorist attacks, National Guard troops stationed at Fort Hamilton provided critical operational

  3. #3 peak

    a Research Paper | Pen and the Pad There are two types of research<|endoftext|>The Apollo 11 space flight landed the first humans on Earth’s Moon on July 20 1969. The mission, carried out by the United

  4. #4 peak

    a man who would go on to become a leading member of the National Left Wing Council and editor of that group's Chicago newspaper, The Communist.⏎Following the September 1918 state convention of the Socialist Party of Michigan, a number of

  5. #5 peak

    “weekends” are not recognised.⏎|Sunday||February 05||Kashmir Day||Observed since 1990

  6. #6 peak

    view at the Portland Museum of Art January 24, 2008, through March 23, 2008.⏎The January 20, 2008, edition of The Boston Globe features an

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 would like to know what I would like to know for the future. I want to know what I want to know for the next year. I want to know what I want to know for the next year. So, I want to

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

    1969 to 51 percent in the most recent estimates.⏎|Sources of Income -

  2. #2 peak

    also argue the Internet (<|endoftext|>このトピックには0件の返信が含まれ、1人の参加者がいます。2020年7月6日 at 9:45 AMに Megan Smith さんが最後の更新

  3. #3 peak

    1978)⏎- National Medal of Science (1986); the APA’s Award

  4. #4 peak

    1928 noted that for children taking vitamin D, there were fewer cavities and the cavity

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. L4:342 suppresses
  2. L2:24 suppresses
  3. L5:7 weakly suppresses
  4. L4:76 weakly excites
  5. L3:187 weakly excites
  6. L3:2576 weakly suppresses
  7. L4:1171 weakly excites
  8. L1:699 weakly excites
  9. L3:130 weakly excites
  10. L5:129 weakly excites
  11. L0:1587 weakly suppresses
  12. L5:703 weakly excites
  13. L4:2850 weakly excites
  14. L1:1855 weakly suppresses
  15. L0:856 weakly excites
  16. L5:1923 weakly excites
  17. L5:1939 weakly excites
  18. L2:493 weakly suppresses
  19. L1:1203 weakly excites
  20. L4:330 weakly suppresses
  21. L4:1441 weakly suppresses
  22. L4:2896 weakly excites
  23. L0:2118 weakly excites
  24. L3:637 weakly excites

Feeds later layers

  1. L13:1155 strongly suppresses
  2. L7:101 strongly suppresses
  3. L16:715 strongly suppresses
  4. L8:423 suppresses
  5. L11:173 suppresses
  6. L14:248 suppresses
  7. L17:163 suppresses
  8. L18:1321 suppresses
  9. L13:74 suppresses
  10. L21:498 suppresses
  11. L13:2099 suppresses
  12. L16:1655 suppresses
  13. L16:662 excites
  14. L20:1197 suppresses 'this' pointing at the current document
  15. L9:1430 suppresses
  16. L12:1065 suppresses
  17. L21:278 suppresses
  18. L21:708 suppresses
  19. L21:68 excites
  20. L12:1755 excites
  21. L20:272 suppresses
  22. L17:1121 suppresses
  23. L17:390 suppresses
  24. L15:668 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 -1.33 and -1.291 values between -0.82 and -0.771 values between -0.67 and -0.631 values between -0.63 and -0.583 values between -0.53 and -0.496 values between -0.49 and -0.448 values between -0.44 and -0.3919 values between -0.39 and -0.3424 values between -0.34 and -0.3053 values between -0.30 and -0.2592 values between -0.25 and -0.20153 values between -0.20 and -0.16279 values between -0.16 and -0.11398 values between -0.11 and -0.06574 values between -0.06 and -0.01702 values between -0.01 and 0.03567 values between 0.03 and 0.08411 values between 0.08 and 0.13279 values between 0.13 and 0.17155 values between 0.17 and 0.2294 values between 0.22 and 0.2748 values between 0.27 and 0.3217 values between 0.32 and 0.3611 values between 0.36 and 0.414 values between 0.41 and 0.464 values between 0.46 and 0.503 values between 0.50 and 0.55offnegativepositiveoften

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. жить
  2. idel
  3. pez
  4. <Resource
  5. 莎
  6. フル
  7. affles
  8. unde
  9. [res
  10. mina

Suppresses

  1. 型
  2. ية
  3. 性
  4. 度
  5. 型企业
  6. 性和
  7. 者
  8. type
  9. 关系
  10. s